Prior Pilot Studies Awardees
2018 Faculty Awardees
Assistant Professor of Hematology/Oncology
Predictors of Functional Impairment in Older Adults with Head & Neck Cancer
The proposed study will be evaluating the trajectories of physical function longitudinally in older (age 60+ years) adults undergoing curative-intent chemoradiation for head and neck cancer. This study will be able to better define patients who are more vs. less likely to develop frailty during and after treatment by evaluating sarcopenia (i.e., loss of muscle mass) and other clinical factors by evaluating time-related changes in physical function. As exploratory aims, this study will look at time-related changes in aging-related biomarkers of inflammation and immunosenescence (i.e., changes in T cell subsets) and how they interact with changes in physical function and rates of recovery. Finally, feeding tube utilization and duration will also be examined with respect to these changes.
Research Associate Professor of Obstetrics and Gynecology
Targeting Vitamin D Receptor for Treatment of Pancreatic Cancer
Pancreatic Ductal Adenocarcinoma (PDA) with a 5-year survival rate of ~6% carries the worst prognosis among malignancies. Relapsed PDA have limited therapeutic options and progressive second and third line chemotherapies have drastically diminishing impact on patient survival, besides, debilitating side effects exceedingly degrade PDA patients’ quality of life. Identification of newer targets and their pharmacologic inhibitors can prevent the prevailing trend. We analyzed microarray data of primary PDA patient samples and correlated the gene expression with overall survival from two independent clinical trials using R2 genomics and Visualization platform (R2genome) to observe that Vitamin-D receptor (VDR) is highly overexpressed in PDA and a poor prognosticator of survival. VDR is associated with proto-oncogene KRAS and immune checkpoint inhibitor ligand PD-L1. Similar to VDR, KRAS and PD-L1 are poor prognosticators of PDA. To block the tumorigenesis orchestrated by VDR and its association with KRAS and PD-L1 in PDA, we have developed a novel antagonist MeTC7 that inhibited proliferation and tumor growth of two independent human PDA cell-line derived xenografts. Given that MeTC7 treatment downregulates PD-L1 in tumor cells and macrophages, we seek CTSI’s pilot award to determine the antitumor efficacy and effects on tumor immune milieu of syngeneic PDA models that recapitulate human PDA pathogenesis closely to assess the clinical potential of MeTC7 in PDA treatment.
Professor of Emergency Medicine, Pediatrics and Public Health Sciences
Piloting Interventions to Improve Adherence to Cervical Cancer Screening in ED
Cervical Cancer is among the most preventable forms of cancer, however, cervical cancer screening rates are far below national goals and have been dropping for over a decade. Lower levels of adherence to recommended screening protocols have been particularly pronounced among racial and ethnic minorities, patients with lower education levels, and patients who use the emergency department (ED) for their usual source of care. Because a disproportionate number of patients with the socio-economic factors that are associated with vulnerability for non-adherence with screening recommendations are cared for in the ED, this setting may be an effective environment to improve adherence to screening recommendations. The overall goal of the proposed study is to pilot two brief behavioral interventions aimed at improving adherence to cervical cancer screening among ED patients. A randomized pilot trial will be conducted in order to compare the efficacies of two structured behavioral interventions in promoting adherence of women aged 21 - 65 to U.S. Preventive Services Task Force (USPSTF) cervical cancer screening recommendations: Intervention A - Screening & Referral and Intervention B - Screening brief mobile technology-based intervention grounded in the Theory of Planned Behavior, & referral. In addition the proposed study will identify predictors of adherence and non-adherence to USPSTF recommendations at baseline. A total of 300 women, ages 21 - 65, will be enrolled through the URMC ED, with an expected sample of approximately 60 non-adherent subjects at baseline. Follow-up will occur at 150 days after the intervention and will be facilitated, by the URMC Women's Health Practice. The effects of Intervention A and Intervention B on adherence at follow-up will be compared, and predictors of baseline adherence status will be examined using χ2 analyses, and predictors of baseline and follow-up adherence status will be examined using hierarchical logistic regression. Upon completion of the proposed pilot project, we will refine these novel intervention models and compete for funding from the NIH in order to assess their efficacies in a large scale randomized controlled trial.
2018 Trainee Awardees
Carol Fries Simpson, MD
Fingerprinting the most curable leukemia: a step toward de-escalation of therapy
Understanding the genetic mechanisms most associated with B cell acute lymphoblastic leukemia (B ALL) susceptibility to therapy has the potential to inform appropriate de-escalation of therapy and mitigate unnecessary toxicity. B ALL is a genetically unstable neoplasm known to undergo genome-wide evolution resulting in substantial genetic heterogeneity within a given patient’s disease. This genetic diversity has functional implications on lymphoblast phenotype, with important clinical differences observed among genetically distinct leukemic subclones. The immunoglobulin heavy chain (IGH) locus serves as a model investigation site for tracking B ALL subclone populations due to its known constitutive rearrangements. IGH molecular sequence “fingerprints” unique to each leukemic subclone can be identified by well-established techniques for next generation deep sequencing (NGS) of IGH loci. Using a unique tissue bank composed of serial B ALL specimens from throughout the remission induction phase of treatment, these IGH “fingerprints” can be used to quantitatively track subclone population shifts in response to therapy. With reliable methods for identifying a cell’s subclone of origin and thus its therapy responsiveness, single-cell gene expression profiling (GEP) can be performed on cells of known treatment susceptibility. This will enable a novel investigation for the GEPs most correlated with favorable therapy response in order to provide new insights about the biologic features most predictive of a favorable outcome. A CTSI Pilot Studies Trainee Award would support the critical experiments necessary to show the feasibility of this approach to simultaneously identifying the GEP and IGH sequence “fingerprint” of single B ALL cells. Once established, these methods can be applied to a subsequent K-08 or equivalent proposal aimed to explore whether the most therapy-responsive leukemic subclones have distinct gene expression features contributing to their susceptibility.
Graduate student in the Department of Immunology, Microbiology and Virology
Platelet dysregulation by antiretroviral drugs and HIV-comorbid disease
HIV infected persons currently experience shortened lifespans and increased risk of cardiovascular disease and other inflammation-linked diseases compared to uninfected persons. Importantly, HIV-infected persons are at a higher risk for cardiovascular disease (CVD) than uninfected persons. While antiretroviral drugs (ARVs) can be used to successfully suppress HIV to undetectable levels, ARVs do not cure HIV; as of today, HIV infected persons must be treated with ARVs for life. Unfortunately, ARVs have many known side-effects, and some have been tentatively linked with CVD. HIV-infected persons experience chronic inflammation, evidenced by increased levels of circulating soluble CD40L (sCD40L). Platelets are key mediators of inflammation, but are often overlooked. As platelets are the major source of sCD40L in the circulation, and are in a highly activated state in HIV-infected persons, it is possible that platelet dysregulation is contributing to or driving the chronic inflammation that increases risk of CVD in HIV-infected persons. I hypothesize that
ARVs can dysregulate platelet function, driving them to a pro-inflammatory, pro-thrombotic phenotype and causing chronic inflammation. My unpublished data shows proof-of-principle that ARVs can directly dysregulate human platelet function. Aim 1 will screen for ARVs, alone and in combination as used clinically, for effects on human platelet function in vitro. Recently developed ARVs to be screened include the integrase inhibitors dolutegravir and elvitegravir; protease inhibitors cobicistat, darunavir, and atazanavir; and entry inhibitor maraviroc. Additionally, older but widely used ARVs will also be included in screening. Aim 2 will begin to determine the mechanisms whereby certain ARVs dysregulate human platelets. The ultimate goal in determining the mechanism of dysregulation is the development of new agents or use of an existing pharmaceutical to prevent the negative effects of the ARVs on platelet function.
Ying Meng, PhD, RN, ACNP
Assistant Professor of Nursing
The influence of brain food reward system on the development of obesity
Brain reward system plays a key role in the development of obesity. The reward system provides pleasure when food is consumed, especially highly palatable food such as sweets and high fat diet. People will eat the food they like, even when they are not hungry. The “reward” from a certain food can override normal appetite regulatory signals and result in consuming more calories than needed. Studies have suggested that addiction and obesity share similar abnormality in the brain reward system. However, what genes are involved in the central food reward mechanisms is still unclear. Therefore, the first specific aim of this project is to conduct a large candidate gene study to examine the association between addiction-related genes and body mass index in adults and children, separately. The second aim is to examine the association of these addiction related genes with food eating behavior in a subset of participants. The identification of these genes and their relationship with obesity will further delineate the genetic constitution of the brain food reward system and facilitate health care providers to develop approaches to manage obesity more effectively.
David Paul, MD
Visual cortex plasticity and white matter changes associated with GH and IGFI
The effects of growth hormone (GH) over secretion are far reaching throughout the human body – including endocrine and metabolic disturbances, cardiomyopathy, congestive heart failure, and enlargement of the hands, feet, and frontal bones. In the central nervous system, GH secreting tumors cause compression of the optic chiasm and a well described pattern of vision loss. To date however, no study has looked at the role of increased levels of GH and insulin like growth factor (IGF-I) on the human visual system. Those studies in humans that have provided insight into the role of GH and IGF-I on the human nervous system focus on 1) broader cognitive processes that rely on complex neural networks, and 2) preferentially study the effects of growth hormone deficiency and its subsequent replacement. Additionally, research on visual function in pituitary tumor patients has focused on identifying predictive factors for visual recovery following decompression of the anterior visual pathway. This literature has not factored in the impact of pituitary hormones on cortical plasticity and brain function. We address this gap by studying both the functional and anatomic properties of the early visual cortex in hormonally controlled GH-secreting adenoma patients (cured) as compared with uncontrolled and treatment naïve patients (uncured).
IGF-I has been implicated in modulating physiologic striate cortex plasticity following monocular deprivation in mice and the effects of experience enrichment in visual cortex development. In adult rats, exogenous administration of IGF-I reinstates juvenile-like plasticity, with shifts in ocular dominance columns following adult monocular deprivation, and facilitates the return of visual function in adult amblyopic animals. The role of excess IGF-I and GH on cortical plasticity in the adult human brain, however, is poorly understood. Studies show that pathologic excess of GH from somatotrophic tumors correlates with decreased electrical neuronal activity and impairment of visual memory tasks.
Here, we propose to use a combination of optical coherence tomography (OCT), functional magnetic resonance imaging (fMRI) and cortical thickness analyses to characterize functional and structural changes in early visual cortex as they relate to circulating levels of GH and IGF-1 in cured vs. uncured acromegaly patients. Unique to this research paradigm is our ability to back project both functional and cortical thickness data onto a retinal map for direct retinotopic comparison with measures of retinal ganglion cell health.
2018 Novel Biostatistical and Epidemiologic Methods Awardees
Assistant Professor of Biostatistics and Computational Biology
Machine learning based mediation analysis: Application in a study of birth weight
A mediator is a variable that lies on the causal pathway between the treatment and outcome and is associated with both the treatment and the outcome. Mediation analysis seeks to explain the role of causal mechanisms that transmit the treatment effect. The existing methods with multiple mediators on the causal pathway from policy to the outcome either assume that the list of true mediators is known a priori, or include all the available mediators in the analysis. The latter approach is recommended to avoid biases caused by ignoring true mediators, but it can potentially inflate the standard error of estimated parameters. Another limitation of existing mediation methods is that linear models are commonly fit and evaluated using linear regression, and thus are susceptible to model misspecification.
The primary objective of this proposal is to develop a methodology that can be used for selecting important mediators, and to improve model fit by leveraging machine learning techniques. Mediation analysis involves two models; one for the outcome and one for the mediators. Standard variable selection methods such as LASSO, which is applied to either of these models separately, may cause some important mediators not to be selected. In Aim 1, we will generalize the LASSO such that the proposed method incorporates both the outcome-mediator and mediator-treatment association in the selection procedure. This can notably improve the efficiency of the estimated parameters while maintaining their unbiasedness. Then we will use a flexible machine learning approach (e.g., random forest), to model the association between the outcome (or mediators) with the confounders, which should improve model prediction relative to linear regression. The main challenge will be that such flexible models do not have the desired root-n convergence rate so they can be very slow to convergence. To handle this, in Aim 2, we will use a procedure that guarantees root-n convergence of the parameters as long as the outcome and mediator models have convergence rate faster than !"/$. That rate is achievable by many machine learning methods including random forest. This aim will considerably reduce the chance of model misspecification in mediation analysis.
Associate Professor of Biostatistics and Computational Biology
Personalized Medical Image Analysis Based on Partial Differential Equations
Medical images, compared with other types of clinical or bio-assay data, are often considered as “semiquantitative” because it is very challenging to study them with conventional statistical methods. The standard practice typically rely on averaging image characteristics over regions of interests (ROIs) defined a priori. Subsequent statistical analyses based on these summary statistics can discover common abnormal patterns shared by most subjects in the study but not subject-specific changes over time.
Recently we developed a medical image analysis pipeline, dubbed as the improved spatial regression analysis of diffusion tensor images (iSPREAD), with the following advantages: a) it does not depend on a priori defined
ROIs; b) it is designed for detecting subject-specific longitudinal changes/effects; c) it uses nonparametric techniques so its validity does not depend on normality; and d) the brain image is represented as a smooth spatial function governed by the Perona-Malik partial differential equation (PDE), which is resolution independent.
However, iSPREAD needs much computational resources because we have to solve the Perona-Malik equation for thousands of times. We propose to reduce computational cost by replacing the finite difference PDE solver in iSPREAD by the finite element method (FEM), which is a more advanced and highly efficient PDE solver that have been used in engineering for several decades. Once developed, we will apply the new method retrospectively to brain images collected from multiple sclerosis patients. Using these data as well as carefully designed simulations, we will be able to study the optimum tuning parameters, and thoroughly compare the performance of the proposed method with its competitors.
Furthermore, we propose to adapt the best PDE technique, the finite element method, for high-dimensional functional data analysis (FDA). Due to the use of weak formulation, FEM-based PDE solutions can be represented as piecewise spatial linear functions. The inner product, roughness penalty, etc, can be efficiently computed based on linear algebra techniques. This is novel and could revolutionize methodological research in multi-dimensional FDA.
2018 Community-Based Participatory Research Pipeline-to-Pilot Awardees
Growing a Healthy Community: Families Co-Constructing Community Spaces and Sustainable Access to Fresh Food
Building on our community-based participatory research practices and relationships in the Beechwood community, this pilot study is designed to inform our understanding of the impacts of food insecurity, local practices that support food well-being, and the active co-construction of gathering spaces. In this urban neighborhood, disparities in health outcomes and access to healthy foods are well documented. Rates of premature mortality within our zip code are “so profound” that all racial and ethnic groups experience significantly worse health outcomes compared with bordering neighborhoods. However, this community is also transforming through the efforts of engaged community members and activists. One current initiative, the Beechwood Greenhouse Collaborative (BGC), is being developed by a team of community families,
university faculty and students to address food access and health inequities through creating sustainable gardens and gathering spaces and is grounded in the intersections we find between food, community and family well-being.
Food well-being, shaped by economic and built environments and food policies, is documented as a key contributor to individual, neighborhood and societal well-being. In addition, research demonstrates the potential of gardens and gathering spaces to facilitate neighborhood development and health promotion.
Health outcomes include improved cardio-vascular health, mental health, and social well-being. In low income urban neighborhoods, gardens provide spaces for residents to socialize, share resources and organize. Researchers propose that creating community managed and owned spaces through urban agriculture have political, social and ecological implications.
This pilot study is designed to explore and document the potential of local community practices and initiatives to support food well-being and co-construct gardens and gathering spaces. This grant will support the sustained collaborative engagement of a diverse team of community-university researchers and activists to address the questions: How are families accessing food in this neighborhood? What does it mean to a neighborhood to have access to fresh produce? How do community members co-construct spaces that support healthy and sustainable change? Engaging in iterative cycles of co-researching and co-implementing that guide our CBPR practices, we will gather qualitative and quantitative data that allow us to develop both research and implementation designs situated in the challenges and strengths of our neighborhood.
Our proposed research team includes faculty in human development and a director of a local community based organization as Co-PI’s, two graduate students, and two participants of the ENGOAL program (Engaging Older Adult Learners as Health Researchers). In addition to actively collecting data, all research team members will engage in iterative cycles of data analysis at bi-weekly team meetings, will be trained to use software to assist in coding and analysis, and will assist with ongoing dissemination of findings. This pilot will provide the necessary documentation and data to apply for a USDA Community Food Grant and/or a RWJF Evidence for Action Grant. Our research will also address local issues of social and economic justice by providing evidence to inform the design and implementation of sustainable access to fresh foods and community spaces through the BGC, local corner stores, and across the Beechwood community.
Community centered prenatal oral health program
Despite the well-linked association between poor maternal oral health and adverse birth outcomes (particularly preterm and low birthweight infants), and a strong correlation between maternal tooth decay and increased tooth decay in children, many mothers-to-be do not receive timely prenatal oral health care. More than 43% of US pregnant women have not had a dental checkup while 76% admitted to suffering from oral health problems during pregnancy. Moreover, the utilization of oral health care during pregnancy is lower among black women, ethnic minorities and women with socioeconomic disadvantages.
Three main factors contribute to inadequate prenatal oral health care utilization among minority and socioeconomically-disadvantaged women: 1) lack of dentists who provide prenatal dental care; 2) lack of awareness and knowledge of the importance of prenatal oral health care, particularly among women from low income neighborhoods; and 3) real and perceived barriers to utilizing oral health care for community pregnant women. To partially address the first factor, we initiated a prenatal dental clinic at the University of Rochester Eastman Institute for Oral Health (UR-EIOH) in March 2018. Few sustained efforts or studies, however, have addressed the second and third factors.
Our long-term goal is to develop an innovative community-centered prenatal oral health program, particularly for socioeconomically-disadvantaged women, that improves access to prenatal oral care, eliminates or reduces dental emergencies during pregnancy, reduces prenatal oral diseases related to adverse birth outcomes, and ultimately improves oral health in children. This community based participatory research proposal will build a partnership between the Heathy Baby Network (HBN) and UR-EIOH and UR-Family Medicine (URFM).
By the end of this CBPR Pipeline-to-Pilot grant, we will have developed a network of care providers and established a robust CBPR committee; assessed the perceived need for oral health care; identified real and perceived barriers to oral health care; developed community-centered strategies to promote oral health during pregnancy for women residing in the most impoverished and least served areas of Rochester; and we will have developed a pilot CTSI, NIH or PICORI proposal to address the most promising approaches to further eliminate or reduce oral health disparities in this group.
2018 UNYTE Translational Research Network Pipeline-to-Pilot Awardees
Developing a virtual reality approach to study and rehabilitate vision after stroke
Stroke-induced occipital damage is an increasingly prevalent, debilitating cause of partial blindness, which afflicts about 1% of the population over age 50. Until recently, this condition was considered permanent. Over the last 10 years, the Huxlin lab has pioneered a new method of retraining visual discriminations in cortically blind (CB) fields. Most recently, we have used this new method to locally recover a range of visual abilities in previously blind fields, even months to years after the original stroke. Although exciting and life altering for patients, and a game-changer clinically, our approach faces some major limitations:
(1) For training to work, stimuli must be presented repeatedly at the same blind field locations relative to fixation. During in-lab testing, we use a binocular eye tracker (Eyelink 1000) to enforce this. At home, where our patients do most of their training (300-600 trials/day, 5-7 days/week for a minimum of 3-6 months), we simply “trust” them to fixate correctly during stimulus presentation. Only about 60% of patients are capable of doing this correctly when unmonitored.
(2) To deliver the stimulus to the target blind regions, patients are required to fixate centrally during peripheral stimulus presentation. This behavior is highly unnatural, leading us to question if it is the most efficient way to achieve perceptual learning. VR could allow dynamic, gaze-contingent stimulus presentation to deliver stimuli in the blind field during natural behavior and free viewing.
(3) Finally, single-target-on-a-uniform-background stimulus delivery requiring only a perceptual judgment may not be the most effective way to retrain vision in CB. An alternative notion is that additional cues present in the real world (e.g. sound, depth, color and target-in-background context), as well as the need to actively interact with the stimulus, could enhance attentional deployment and/or feature learning with respect to the target, dramatically boosting learning efficacy.
One specific technological development could help us address all these limitations: integrating precision eye tracking into a modern VR helmet, and porting the CB visual retraining paradigm into this system.
This is where collaboration with Rochester’s Institute of Technology (RIT), which is a UNYTE institution, has become necessary. Recently, Dr. Huxlin approached Dr. Diaz, Director of the Perform Lab, located at RIT’s Center for Imaging Science. The Perform Lab houses several highly-ergonomic and high-resolution HTC VIVE virtual reality displays with integrated low-latency eye trackers by SMI and Pupil Labs. Dr. Diaz has over a decade of experience conducting studies of human visually guided behavior in virtual environments, and is becoming a leader in the area of eye tracking in virtual reality. Dr. Diaz sought out two RIT students, who recently began working on this project, integrating the SMI eye-tracker into a HTC VIVE, and demonstrating feasibility of programming Dr. Huxlin’s visual training paradigm into the system. We now seek funds to partially support these students during the next year, so that we can complete the hardware/software development on our system and garner preliminary data of its performance necessary to seed further grant support.
Measuring postpartum contraceptive uptake and interconception periods in the electronic medical record
Population health indicators to improve birth outcomes – including preconception health metrics – can be measured at various points in time, including during the period of time before pregnancy (preconception period) and during the period of time in between pregnancies for women with more than one pregnancy (interconception period). Examples of CDC and Healthy People 2020 indicators to improve preconception and interconception health include markers such as the proportion of deliveries to women who used contraception to plan their pregnancies and proportion of women with adequate birth spacing, defined as 18 months between birth and subsequent pregnancy. Evidence supporting the use of these indicators to improve birth outcomes is well-known. Electronic medical record (EMR) documentation of obstetrics and gynecology care provision is typically set up to focus on each pregnancy as a separate event for female patients. Therefore, EMR data does not easily lend itself to quick referral to data from the first pregnancy and comparing it to the interconception period and, then, subsequent pregnancy. Examples of helpful clinical information across time include determining how decision-making around use of postpartum contraception supports birth spacing goals for patients across pregnancies, and if clinical birth spacing conversations are linked to desirable and recommended interconception periods. This project focuses on two related outcomes: postpartum contraceptive use and birth spacing. Both of these outcomes are constructed of two components: documentation of patient-provider conversations about each topic and quantitative data for measurement. This project will: 1) calculate baseline percentages of each outcome and create an automated EMR calculation module showing histories of postpartum contraception and birth spacing conversations plus the amount of time since last birth in two different UNYTE institutions, 2) compare the two EMR systems, and 3) compare the ease of mining data and creating an automated EMR module. The study population includes URMC Women’s Health Practice patients who gave birth at Strong Memorial Hospital and all UHCC Women’s Health Clinic patients who gave birth at Crouse Hospital between November 1st, 2013 and September 30th, 2017. Results will immediately be used to positively impact clinical care surrounding interconception periods. Results will also influence the development of further research – a patient-provider communication intervention that seeks to improve postpartum contraceptive planning and supports birth spacing conversations while simultaneously improving patients’ decision-making ability concerning postpartum contraceptive methods.
2017 Faculty Awardees
Assistant Professor of Medicine in the Aab Cardiovascular Research Institute
Cardiac Biomarkers for Sudden Unexpected Death in Epilepsy
1% of the US have epilepsy, and those with epilepsy have a 24-fold higher risk of sudden death. The cause of death is often unknown and thus, termed Sudden Unexpected Death in Epilepsy (SUDEP). Indirect evidence links SUDEP to seizure-induced respiratory dysfunction, dysregulation of brain circulation, and cardiac arrhythmias. Cardiac arrhythmias have been recorded in patients and animals with epilepsy immediately preceding sudden death. Many genetic forms of epilepsy, particularly SUDEP cases, are linked to mutations in genes expressed in both the brain and heart. I have shown it is critical to take a MULTI-SYSTEM APPROACH to understand the mechanisms for electrical disturbances in both hearts (arrhythmias) and brains (seizures) of patients with epilepsy.
Detailed ECG analyses in patients at risk of SUDEP aligns with my LONG-TERM GOAL to take a multi-system and translational approach to: (1) elucidate the mechanisms for neuro-cardiac electrical pathologies in epilepsy and (2) develop strategies to predict and reduce SUDEP. The NIH funded Center for SUDEP Research collected 1282 multi-day patient recordings of simultaneous electrical activity in the heart (ECG) and brain (EEG), and measures including respiration, blood pressure, and video monitoring. While there are reports of ECG pathologies and arrhythmias in epilepsy, the results are mixed and the prevalence, especially surrounding seizures, are not fully understood. We HYPOTHESIZE ECG abnormalities and arrhythmias, particularly around times of seizure activity, serve as non-invasive cardiac biomarkers of future SUDEP. We will perform standard ECG measurements and static/dynamic ECG measurements, which are associated with autonomic regulation. AIM 1: Investigate the relationship between baseline ECG pathologies vs. a history of epilepsy or risk of future SUDEP. AIM 2: Assess the incidence, dynamics, and types of cardiac ECG pathologies surrounding seizures/events.
We can't predict SUDEP since we don't know the mechanisms. This study will use the first systematic collection of 1282 patient recordings and >500 clinical parameters to perform the most comprehensive analyses of ECG measures in epilepsy. Results will increase our understanding of the neuro-cardiac pathologies in epilepsy, particularly in those at the highest risk of SUDEP. We will identify robust predictors for SUDEP.
Assistant Professor of Urology
Patient-specific simulated procedure rehearsal for minimally invasive surgery
Simulation has provided opportunities for trainee and expert operators to practice specific procedures, or related skills, prior to performing complex tasks on patients. Although the benefits of simulation training have been well established, other methods to utilize simulation technology remain under explored. Evidence suggests that simulation immediately before criterion surgical tasks (preoperative simulation) benefits performance. However the type of simulation (rehearsal or warm-up) that best enhances surgical performance has not been established. Virtual reality trainers provide a platform for surgeons to refine their skills using generic tasks that are not procedure specific or patient specific. Recent advances in 3D printing of biologic material, coupled with software that incorporates imaging data into a computer design, make it possible to develop individualized models from patient imaging data. In this proposal we aim to assess the impact of generic versus patient specific preoperative simulation platforms on operative performance and patient outcomes in both experts and trainees.
In a randomized fashion, 8 urologists (5 experts, 2 residents and 1 fellow) will perform 6 patient specific simulated rehearsal (PSR) simulations and 6 virtual reality simulation, warm-up (VRS) simulations, using stratified sampling prior to completing a minimally invasive partial nephrectomy on 12 patients with localized renal cancer. To assess technical performances both subjective (global evaluative assessment of robotic skills) and objective (blood loss, operative time, ischemia time, positive margins) performance metrics will be measured. Patient outcomes will be assessed by improvement in hospital stay, complications & decline in renal function. Additional information will include trainees’ autonomy and surgeon takeover.
Through this work we could identify which types of simulation and methodological approaches stand the best chance of demonstrating a convincing link between performance in the simulation and patient safety outcomes. Furthermore, preoperative rehearsals can be used as a "check-out" procedure (similar to the "check ride" used for newly minted aviators) to ensure clinician proficiency prior to performing intricate, risky procedures on actual patients.
Professor of Public Health Sciences
Developing Social-Network Measures of Medical Staff Interactions in Nursing Homes
Increasingly, nursing homes (NHs) have become the place of post-acute and long-term care for over 3.5 million older and frail persons annually. Although the majority of nursing home care is to provide the support of basic functions of daily living, the underlying causes for these functional limitations are addressed by clinicians who diagnose, treat and monitor the medical conditions. In light of a growing shortage of primary care physicians specially trained in geriatrics and/or committed to NH care, the need to identify structures that optimize physician practice and enhance quality within the NH becomes even more pronounce.
Coordination of care through medical staff exchange of clinical information, care referrals and shared learning are increasingly understood as essential building blocks of excellent medical care. 5 Social Networks
Analysis (SNA) provides tools to study interpersonal NH medical staff interactions. Measures of these networks can describe the organization of medical care provided in a nursing home. Similar to other successes with SNA, this line or research is essential to the development of management tools to assist nursing home administrators and medical directors in creating medical staff teams that are capable of providing the best care to nursing home residents. We propose to use a unique dataset that includes surveys of about 2,000 NHs and complete Medicare claims, enrollment information and resident assessments in those NHs that is linked using the Residential History File methodology 9 which allows locating Medical Staff visits in particular NHs. We will use this dataset to develop nursing home specific social networks and derive network measures that correspond to surveyed concepts. The overarching goal of this line of research is to study determinants and consequences of NH medical staff organization as measured on the population of
NHs over time in order to determine NH specific optimal medical staff organization. The specific aims of this developmental grant are to obtain initial results that would be used to develop a line of National Institute of
Aging (NIA) funded research using administrative data. Specifically, the aims of this one-year CTSI sponsored project are: 1) Develop and characterize social network structure in nursing homes in terms of
SNA measures relevant to NH care and determine which measures or combination of measures best captures survey measures of nursing homes medical staff organization; and 2) Explore the relationship between the SNA measures developed in Aim 1 and NH’s quality of care in terms of 30-day rehospitalizations and prevalence of use of antipsychotic medications.
Associate Professor of Pediatrics and Otolaryngology
Sensitivity to Envelope Structure in Children with Otitis Media and Hearing Loss
Conductive hearing loss due to otitis media is common in children during critical periods of auditory and speech development, but the long-term consequences of transient otitis media-related hearing loss on auditory perception are unclear. Previous clinical studies have produced conflicting results, but these studies included subjects with widely varying degrees and durations of hearing loss, and focused largely on higher order processing abilities. Studies of fundamental auditory processes, such as temporal envelope sensitivity, in subjects with a history of measurable conductive hearing loss might produce more consistent results. Envelope processing plays a key role in speech perception, but very little is known about the early development of envelope processing, and whether temporary impairments in hearing during infancy and childhood can affect envelope perception later in life. The purpose of this research is to better understand how a history of otitis media-related hearing loss affects the ability of children to detect and process fluctuations in complex sounds, and what strategies and cues they may use to detect sounds in noise. The aims of our research are first, to determine if amplitude modulation (AM) detection is impaired in children with a history of otitis media-related hearing loss, after restoration of normal hearing; and second, to determine if this early hearing loss affects detection of tones in noise later in childhood. The third aim will use a computational model of envelope processing in the midbrain to explore the effects of reduced central inhibition (a candidate mechanism for behavioral deficits) on model neural sensitivity to the same stimuli. We hypothesize that temporary, measurable conductive hearing loss due to otitis media in otherwise healthy children between the ages of 6 months and
3 years will result in deficits in behavioral sensitivity to envelope fluctuations that persist into later childhood, after restoration of normal hearing. In the proposed study, children between the ages of 4 and 7 years with normal hearing but either 1) a documented history of otitis media and conductive hearing loss from age 6 months to 3 years, or 2) no otitis media or hearing loss, will undergo audiologic examination of both ears followed by experimental sessions during which AM detection thresholds and tone-in-noise thresholds are determined. Modelling results will show whether diminished inhibition in midbrain sensitivity to these stimuli correlate with behavioral deficits in children. Our preliminary data from 15 children, primarily with a history of otitis media-related conductive hearing loss, show significant maturation of AM detection thresholds from age 4 to 7, with older children achieving greater AM sensitivity for low modulation frequencies.
2017 Trainee Awardees
Graduate student in the Department of Genetics, Development and Stem Cells
Using Gene Networks to Translate Bone Quality into Novel Osteoporosis Therapies
Osteoporosis remains a significant concern. By 2025, osteoporotic fractures are expected to increase to 3 million with an estimated cost of $25.3 billion. Morbidity and mortality increases following all major fractures in patients over 55 years of age. It is widely understood that compromised bone strength is the underlying pathophysiology of osteoporosis. Currently, bone mineral density (BMD), is the basis for diagnosis and the main target for osteoporosis therapy. BMD is an important clinical tool, but explains only 55 % of fractures, leaving the needs of 45% of patients unaddressed. Fundamentally, bone strength has two interconnected but distinct components: quantity and quality. While BMD reflects quantity, bone quality (BQ) reflects morphologic and compositional properties. However, developing therapies based on BQ is limited by two major gaps in knowledge: 1) Which genes regulate BQ and 2) Does estrogen deficiency modify the genetic regulation of BQ? Therefore, the objective of the work proposed is to construct a genetic network to identify candidate genes regulating bone composition (SA1) and determine if estrogen deficiency interacts with these genes(SA2).The rationale for the proposed work is that bridging these gaps in knowledge is a crucial first step in the identification of novel therapeutics. This study is responsive to the Funding Opportunity in that it addresses a significant public health issue, capitalizes on the multidisciplinary expertise and offers exceptional training potential for the Trainee (PI: Mr. Beltejar), and has the strong potential to catalyze future research. The results produced from SA1 will inform ongoing research in the lab--in particular, a genome-wide association study looking for locations in the genome that are responsible for bone quality. Secondly the results of SA2 have a strong potential to unlock research on multiple levels osteoporosis management. Identification of novel genes could be used as a new biomarker to identify at risk individuals earlier. Downstream activity of these genes could become additional metrics to monitor disease progression. Perhaps, the most apparent of which is that these genes become putative targets for novel therapies. The research in this application is innovative because it is a significant departure which shifts focus from BMD to BQ as an equally important contributor of fracture resistance.
Cardiovascular Disease Fellow
Patient-reported Outcomes for Patients Undergoing Aortic Valve Replacement
Aortic stenosis is one of the most common cardiovascular conditions in the United States. The treatment approach to aortic stenosis is rapidly evolving as TAVR has been developed. TAVR is indicated for patients at high risk for death or complication with SAVR. The number of TAVR procedures being performed will vastly increase as this treatment option becomes an option for patients at low to intermediate surgical risk.
Patients will complete survey instruments on pain, physical functioning, and mental health prior to their appointment and/or procedure. These instruments were developed by the National Institutes of Health as part of the Patient Reported Outcomes Measurement Information System program (NIH PROMIS). We will identify a cohort of patients presenting for TAVR using the valve program lab clinical databases, and will retrospectively collect the patient’s quality of life data from the medical record. Our measurement system is operational, and preliminary data is available. Our capture rates currently exceed 90%. Within this cohort, we will examine the association between patient-reported health status at baseline and clinical decision making. We will also examine differences in the decision to perform valve replacement, and allocation of treatment TAVR vs. surgical). We will also examine whether outcomes differ between surgical or TAVR valves.
2017 Novel Biostatistical and Epidemiologic Methods Awardees
Assistant Professor of Biostatistics and Biomedical Genetics
Estimation of cell-type specific microRNA expression in complex tissue samples
Tissue samples are not homogeneous entities; rather they are comprised of multiple cell types with distinct functions and corresponding transcriptomic profiles. The proportion of component cell types in a sample often varies from sample to sample and from person to person. In fact, early investigation of gene expression in blood concluded that a significant amount of inter-individual variability in expression could be traced to differences in the proportion of cell types. Furthermore, many pathological processes alter cellular composition via mechanisms such as infiltration or differentiation. This finding has been used to explain the irreproducibility of several serum/plasma miRNA biomarker studies, which highlights the importance of modeling cellular composition before attempting to elucidate pathobiological expression alterations. Analysis of complex tissue samples without accounting for differences in cellular composition can lead to erroneous conclusions such as the design of clinical interventions targeting a microRNA that is unexpressed in the diseased cell type.
The overall goals of the proposed research are to develop statistical deconvolution methodology to estimate the cellular composition and cell-type specific microRNA expression of complex tissue samples and to apply this methodology to analysis of ~20,000
Assistant Professor of Neurology in the Center for Health and Technology
Development of a Clinical Trial Simulation Tool for Huntington's Disease
The objective of this proposed Novel Biostatistical and Epidemiologic Methods project will be to develop disease progression models of Huntington’s disease (HD) that will serve as a backbone to a clinical trial simulation research tool. Besides drug inefficacy, failures in neurodegenerative diseases like HD may be attributable to insufficient trial design including inadequate dosing, population selection, or poor design optimization (e.g. poor choice in sensitive end points, or statistical methods. Given the unmet medical needs and costs associated with the development of neurodegenerative therapies, it is important that the unnecessary failure of clinical trials attributed to design factors be minimized as much as possible. Clinical trial simulations allow for an investigation of the impact of a range of design characteristics on the likelihood of detecting a treatment effect before actually carrying out the study and exposing individuals to investigational agents. A clinical trial simulation tool has three basic modeling components: drug effects, disease progression, and clinical trial design features. Thus, a prerequisite for clinical trial simulation is to have models that allow for the simulation of the time course of disease symptoms or markers in individuals (i.e. disease progression models). The immediate goal of this project is therefore to model the longitudinal progression of clinical outcome measures of HD using data from clinical trials and longitudinal observational studies available at the Center for Human Experimental Therapeutics (CHET).
2017 Community-Based Participatory Research Pipeline-to-Pilot Awardee
Many members, one body: Integrating systems in mental health and self-injury resilience
The proposed project “Many members, one body: Integrating systems in mental health and self-injury resilience,” seeks to build systems science informed approaches to optimize faith communities’ response to pressing mental or behavioral health burdens and related self-injury prevention needs in their congregations and surrounding communities. Working collaboratively with clergy and lay leaders, we will conduct and study participatory group model building and system activation activities that include community training in (1) system thinking and behavioral health topics to increase mental health related knowledge, skills, attitudes and supportive relationships (e.g., with care settings, etc.). Using multi-method approaches, we will examine how faith communities, especially in congregations serving African Americans, develop their system insights and deepen their engagement in mental health via a collective impact approach to change the culture of health and advance their community resilience. Effect sizes will be examined for the relationship of developing mental health promotion programming and partnerships on stigma reduction and action engagement aimed at systemic approaches to reducing and preventing mental health morbidity and related self-injury mortality in the Finger Lakes region of New York State surrounding the city of Rochester, New York.
The lead partners bring together a history of excellent collaboration, a firm dedication to (and prior success at) working with underserved populations, and unique strengths that will inform the various aspects of this project. Ann Marie White contributes: a) expertise on mental health promotion for emerging adulthood, b) integrating system thinking perspectives and scientific methods in public health and injury prevention, c) national level policy experience, d) community engagement and collaboration in health systems improvement and research, and e) experiences working with “Big Data” (data science) in assessing self-injury risks and protective factors through social media at a community level. Phyllis Jackson contributes her faith-community organizing from working with Interdenominational Health Ministry Coalition. She also has vast direct community health promotion experience as a community nurse, most recently overseeing community engagement that advances blood pressure control at population levels for Common Ground Health. Silvia Sörensen contributes experience designing and evaluating and longitudinal intervention research projects, and a focus on middle-aged and older adult mental health and wellness – from individual and population perspectives.
Together with other community partners in a community-based participatory research approach, this research team will work to develop models that help inform decisions and generate system level insights among faith, prevention (including public health) and health system leaders who seek to work together in a collective impact framework to improve population health in mental and behavioral health related morbidity and mortality. This research builds upon their work together leading Renewing of the Mind (RoM). This program, now in its fifth year is managed in collaboration with faculty and staff from the Department of Psychiatry, Strong Recovery, Trillium Health and Mental Health Association of Rochester. RoM trains area church members and leaders serving members of the African American, Hispanic and Latino communities in the Finger Lakes region to advance behavioral and mental health topics in their faith community settings.
2017 UNYTE Translational Research Network Pipeline-to-Pilot Awardees
Transplacental Transfer of Vitamin D3 and 25(OH)D3 in Human Pregnancy
Suboptimal vitamin D status during pregnancy has been linked to a growing list of adverse maternal and neonatal birth outcomes including increased risk of C-section, preeclampsia, gestational diabetes and small for gestational age. These potential adverse effects of low maternal D status are of concern since 21% of US pregnant women are at risk for inadequacy (25(OH)D from 12 to <20 ng/mL) and 7% are at risk of D deficiency (<12 ng/mL). The fetus is entirely dependent on maternal D status and D insufficient women will give birth to D insufficient neonates. At birth umbilical cord concentrations of 25(OH)D in the neonate are 20-30% lower than maternal concentrations but at this time the mechanism by which vitamin D is transferred across the placenta is unknown. Existing dogma posits that maternal 25(OH)D, but not the parent vitamin (D3) or the hormonal form of vitamin D (1,25(OH)2D), is transferred to the fetus based on a lack of correlation between maternal 1,25(OH)2D and neonatal 1,25(OH)2D. A recent paper published only this month challenges many of the assumptions in the existing literature as this study found that fetal, and not maternal D genotype, was significantly associated with neonatal D status at birth. We recently used CTRA pilot funding to successfully develop a deuterated cholecalciferol method that allowed us to measure the absorption of D3, its conversion into 25(OH)D3 and the half-life of 25(OH)D3 in pregnant (mid-gestation) and non-pregnant women. This was the first study to measure absorption and metabolism of vitamin D in any human population using deuterated cholecalciferol. These pilot data were used for an NIH R01 submission to further characterize vitamin D dynamics during pregnancy, this application is currently under review. In this new pilot grant, we would like to capitalize on our prior findings and use our newly developed method to move this field one step forward to now begin to address the dynamics of maternal transfer of D to the fetus. Dr. Pressman and O’Brien previously undertook a similar placental transfer study of calcium at the University of Rochester using stable calcium isotopes administered to teens at delivery.1 We now wish to employ a similar approach to evaluate maternal-fetal transfer of vitamin D. Three groups of women will be studied; 1) women (n=2) will be dosed with deuterated cholecalciferol in late gestation to allow time for both the mother and the fetus to further utilize or metabolize the parent D3 into 25(OH)D3. Two other groups of women will be studied at delivery and will receive either 2) deuterated D3 (n=4) or 3) deuterated 25(OH)D3 (n=4) to see if these forms of D can rapidly cross the human placenta. Our group is uniquely suited to rapidly address these questions using our recent CTSI funded pilot data that developed and validated UHPLC-MS/MS method. The proposed study is novel and will utilize the interdisciplinary skills of UNYTE investigators that have a history of collaboration and whose expertise spans the fields of obstetrics, D physiology, biomedical mass spectrometry and maternal and child nutrition.
Advanced Digital Stethoscope - Pilot Study of Acoustic Diagnostics for Left Ventricular Assist Devices
Left ventricular assist devices (LVAD) are becoming an ever more important part of the management of end-stage heart failure. LVADs pump blood from the left ventricle to the aorta, pressuring the aorta and providing needed cardiac output to patients with severely reduced heart function. LVAD dysfunction often presents with complaints of dyspnea and fatigue, but there are many other nondevice reasons for end-stage heart failure patients to have these symptoms. We believe that the combination of sounds from the native heart and the implanted LVAD may have an important role in diagnosing patients with suspected device dysfunction. Stethoscope design has changed little since its invention in 1816 by Rene Laennec. We propose to modernize stethoscope design using innovative digital signal processing techniques. The techniques to improve the stethoscope’s acoustic signal include digital signal filtering, advanced beat-based rejection algorithms, and ensemble averaging. The techniques to improve acoustic diagnostics include spectral analysis, advanced automated neural networks and a combination of improved acoustics and smartphone-based interactive software that will allow the clinician to make an informed bedside diagnosis when integrating the advanced acoustic analyses with other routine clinical information. Our goal for this pilot study is to collect the preliminary data necessary for a grant application to fully develop and test an advanced digital stethoscope intended to improve the bedside assessment of patients with implanted LVADs. Over the 6-month period of this pilot study, we will collect simultaneous electrocardiographic (ECG) and acoustic data from patients with an implanted LVAD using standard clinical equipment that has been modified for digital signal acquisition, and we will also test a design candidate for the housing of an eventual monolithic advanced digital stethoscope. We will use already developed acoustic analysis routines and test new ones compared to nearly simultaneously acquired echocardiographic data. Subjects will be recruited from the echocardiography laboratory, so full quantitative ultrasound information will be available at no extra cost to the investigators. Clinician-scientists at University of Rochester Medical Center will collaborate with engineering professors and students at the Rochester Institute of Technology to create an entirely new paradigm of acoustic diagnostics by building on the original stethoscope design, which has not changed appreciably since it was invented 200 years ago. LVAD patients were chosen as the first testing ground for these innovative acoustic diagnostics methods, as this patient population’s unique combination of native and mechanical heart sounds make traditional stethoscope use particularly problematic. The collaboration between the URMC cardiovascular service and RIT Kate Gleason College of Engineering embodies the UNYTE vision of translational innovate research.
2016 Faculty Awardees
Research Assistant Professor of Hematology/Oncology
Targeting of small molecule inhibitors for the treatment of myelogenous leukemia
Acute myelogenous leukemia (AML) is a disease of the bone marrow that is newly diagnosed in over 18,000 patients in the United States yearly with only approximately 25% of patients surviving beyond 5 years.
Morbidity and mortality in this disease are caused, in large part, by loss of normal blood cell development. Hematopoietic stem cells (HSCs) are responsible for the maintenance of the blood system over the life of an individual and rely on a niche of multiple cellular components within the bone marrow microenvironment (BMME) for regulation of cell fate. We and others previously demonstrated that the chemokine CCL3 (Mip1α) was found to be highly upregulated and necessary for BMME modulation and leukemia development in AML. Upon treatment of our murine model of AML with inhibitors of two CCL3 receptors, bx471 and maraviroc, we demonstrated an incomplete improvement in HSC dysfunction. Subpar pharmacodynamics is likely due to inadequate inhibitor half-lives and poor retention within the BMME. To address these hurdles, we hypothesize that targeted drug delivery approaches for inhibitors of the CCL3 pathway to the bone marrow microenvironment will improve pharmacodynamics, realizing therapeutic efficacy. To test this hypothesis, we propose a multidisciplinary approach using small molecules already tested in the Frisch lab loaded into nanoparticles developed by the Benoit lab that target the BMME and result in delayed release, improving retention time in the marrow. We propose to test this drug delivery approach using 2 murine models of AML, as well as xenograft models of 5 de-identified primary patient samples of de novo AML.
Research Assistant Professor of Medicine in the Aab Cardiovascular Research Institute
Effect of Statin therapy in combination with QT prolonging drugs
Our preliminary clinical data in Long QT syndrome patients suggest that statins may be detrimental to patients with mutations in the KCNH2 gene, increasing their cardiac risk. The main goal of this project is to understand the mechanism of action of statins in combination with naturally occurring mutations associated with Long QT syndrome and/or Long QT prolonging drugs that target the IKr channel. Understanding the molecular and whole heart effect of statins will allow tailoring of therapy to patients who would benefit the most and to avoid drug combinations or particular patient population for which statins may be harmful. For that we will first test the hypothesis that statins can exacerbate the QT prolonging effect of arrhythmogenic drugs when given in combination in a rabbit model. We will measure QT prolongation and T-wave morphology changes in ECGs in rabbits. Animals will be treated with QT prolonging drugs (IKr inhibiting β- blockers: propranolol and sotalol) and statins (either simvastatin, atorvastatin and fluvastatin). These statins have different degrees of Cytochrome P450 isoenzyme degradation and therefore different potential degrees of risk of medication interaction when given in combination with QT prolonging drugs. Second, we will investigate the celular mechanism of the underlying effect. We will test the hypothesis that statin potentiation of drug inhibition of the potassium channel IKr explains the adverse drug combination. For that we will measure the effect of statins in combination with the β-blockers propranolol and sotalol on the four main repolarizing cardiac currents, IKr, IKs, ICaL and INaL. In addition, we will test the effect of statin treatment in mutant ion channels linked to long QT syndrome type 2. Finally, we will use a human cardiac computer model to study the combined effect of measured ion channel regulation in cellular repolarization.
Research Assistant Professor of Obstetrics and Gynecology
MeTC7 a novel Vitamin-D receptor antagonist for immunotherapy of ovarian cancer
Epithelial Ovarian Cancer (EOC) causes highest mortality among the gynecologic cancers due to recurrence and chemoresistance. New therapies are urgently needed. Following dose-limiting lethal hypercalemia and hypercalciuria arising from vitamin-D receptor (VDR) agonists such as Calcitriol treatment in clinical trials, overexpression of VDR and its intriguing association with immune checkpoint inhibitor PD-L1 and other immune suppressors such as platelets in ovarian tumors, we hypothesized that VDR antagonists will provide optimal antitumor efficacy and safety. Validating this hypothesis, we developed the first pharmacologically pure and non-hypercalcemic VDR antagonist (MeTC7) that exhibited promising antitumor efficacy in ovarian cancer, melanoma and medulloblastoma xenograft models without causing hypercalcemia. This study aims to investigate whether targeting VDR can lead to antitumor effects in syngeneic ovarian cancer models and restore immune functions or break immune tolerance in ovarian tumors. Being highly translational and team oriented, this study meets the goals of CTSI pilot program by bringing together cross-disciplinary expertizes in cancer biology, medicinal chemistry and immunology in team with a gynecologic oncologist (Dr. Moore) and computational biologist Dr. John Ashton to develop MeTC7 for treatment of EOC, a disease with significant unmet medical needs. If the goals of this study are achieved, a novel small molecule immunotherapeutic agent targeting VDR without hypercalcemic liabilities for treatment of immunogenic tumors may emerge.
2016 Trainee Awardees
Parker Duffney, PhD
Graduate student in the Department of Toxicology
Cigarette Smoke Impairs the Anti-Viral Response to Influenza A Virus
Cigarette Smoke Impairs the Anti-Viral Response to Influenza A Virus: Potential for Resolution-based
Therapies: Over 1 billion people worldwide smoke cigarettes. Epidemiological evidence links secondhand cigarette smoke exposure to increased viral infections in children. Similarly, cigarette smokers are at increased risk for viral infection in the lung that are often accompanied by very severe symptoms. In people with underlying inflammatory lung disease, especially chronic obstructive pulmonary disease (COPD), viral infection of the lung triggers exacerbation of disease symptoms leading to the bulk of the morbidity, mortality, and healthcare costs. The mechanism by which cigarette smoke leads to increased incidence and severity of viral influenza (flu) infection is not understood and there remains an unmet need for therapies to prevent and treat exacerbations in COPD patients. The identification of endogenously produced, anti-inflammatory and pro-resolution lipids, termed specialized pro-resolving mediators (SPMs), offers a novel therapeutic strategy to treat inflammatory diseases. However, the potential of SPMs to restore smoke-induced immune dysregulation in response to viral infection is unknown. We have strong supporting data in primary human small airway epithelial cells that cigarette smoke impairs antiviral responses by decreasing production of antiviral interferons and allowing increased viral replication. We are now going to take these results to a preclinical animal model of infection following cigarette smoke exposure to investigate the mechanism of viral susceptibility and new therapies to mitigate the injury. We hypothesize that cigarette smoke exposure leads to more severe influenza infection due to failure to control virus replication and that smoke-induced defects can be reversed with SPM treatment. Aim 1 will investigate the effect of cigarette smoke exposure on influenza infection using an animal model. Aim 2 will evaluate the ability of SPMs to mitigate cigarette smoke induced defects in response to influenza infection.
MD-PhD candidate in Medicine and Translational Biomedical Science
Importance of Axonal Injury and Inflammation in the Acute Phase of Mild TBI
Mild traumatic brain injury (MTBI), or concussion, presents an immense healthcare burden. In addition to having short and medium term quality of life consequences, it is associated with debilitating long-term neurodegeneration. Currently, early symptoms which are used for diagnosis, are difficult to identify objectively in clinical settings, and do not predict which patients will develop chronic sequelae. More objective pathophysiological markers are needed to improve diagnosis, treatment, and prognosis. Many of the early symptoms and signs of MTBI are related to functions controlled by the brainstem and experimental studies suggest specific geometric vulnerability of this region to concussive loading. However, to date, no human observational studies have examined axonal injury in the brainstem in relation to specific pathways related to stereotyped dysfunctions (e.g. oculomotor dysfunction) of MTBI (Gap 1). Another consequence of
MTBI is aberrant neuroinflammation and blood brain barrier disruption followed by brain-native protein escape after injury; these central features could play a key role in chronic sequelae of MTBI. These inflammatory features of MTBI can incite peripheral inflammation. However, while the central (brain) inflammatory perturbations in MTBI are well understood, the state of peripheral inflammation in MTBI patients has not been studied (Gap 2). To address these 2 gaps in the field, we proposed the following: examine DTI-determined axonal injury in brainstem oculomotor pathways of MTBI patients (Aim 1) and determine the difference in select peripheral inflammatory markers between concussed and non-concussed athletes (Aim 2). This study is responsive to the Funding Opportunity in that it addresses an intractable public health issue (MTBI), utilizes multidisciplinary expertise at the university and offers high training potential for the Trainee (Pl: Mr. Hirad). Both aims of this study are innovative and if successful have the potential to shift our understanding of MTBI in its earliest stages. Immediate future efforts will focus on probing for axonal injury in brainstem pathways related to cognition and autonomic function (functions that are usually affected in MTBI). Future studies will include longitudinal research into the resolution/exacerbation of axonal injury and inflammation in MTBI.
2016 Novel Biostatistical and Epidemiologic Methods Awardees
Assistant Professor of Biostatistics and Biomedical Genetics
Development of qPCR methodology for clinical testing
The most widely used clinical gene expression assay in the US, Oncotype DX, uses quantitative real-time PCR (qPCR) to measure gene expression. Recently, microRNA profiling of blood samples has shown the potential to be a minimally invasive screening procedure, and qPCR-based technologies have been developed to simultaneously measure the majority of human microRNAs. The popularity of qPCR for clinical testing is due to higher sensitivity and lower cost than competing technologies. However, the majority of statistical methodology for qPCR data analysis was developed for laboratory experimentation and is not readily applicable in a clinical setting. In particular, improved methods are needed to handle non-detects, those reactions failing to produce a minimum signal. Such methods will substantially improve the accuracy and validity of clinical qPCR testing and facilitate the translation of genomic biomarkers into clinical practice.
Common procedures for handling non-detects introduce substantial bias into estimates of both absolute and differential gene expression. By modeling non-detects as non-random missing data, we greatly reduced this bias. Our approach focused on qPCR data generated in a laboratory setting in which we could borrow information across replicate samples. However, replicate samples are rare in a clinical setting in which patients, and therefore samples, must be analyzed sequentially. Translation of our approach to clinical analyses will require the development of novel methodology that does not rely on the availability of replicates. The development of these methods is crucial for qPCR-based clinical testing to realize its full potential.
The overall goals of the proposed research are: (1) to develop improved methodology to handle non-detects in qPCR data, (2) to develop a single sample version of our methodology for clinical biomarkers, and (3) to assess the applicability of our methods to microRNA transcriptome screening.
Associate Professor of Biostatistics and Computational Biology
Weighted Functional Gene Set Enrichment Analysis for Time-course Transcriptome Studies
Gene set enrichment analyses (GSEA) are powerful inferential methods widely used in genomic research to provide mechanistic insights that can be validated experimentally. Currently available methods have two shortcomings identified in our preliminary study. 1. Large-scale time-course gene expression studies have gained tremendous popularity in recent years, yet only a few GSEA methods can be applied to these data. These limited options either have unrealistic modeling assumptions or do not utilize the time-course information efficiently. 2. Biologically defined gene sets can have remarkable overlaps, partly because many genes can have different biological functions under different conditions. All existing GSEA procedures ignore this “gene overlapping” and over-weight these genes, which reduces their statistical performance significantly, as shown in our preliminary study. To address these issues, we propose a novel GSEA pipeline based on functional data analysis, which is a powerful and flexible statistical framework that has been widely used in time-course data analysis. We will use a functional F-statistic to capture temporal expression pattern for each gene. Next, we will apply a weighting method based on penalized functional regression to address the gene-overlapping issue. These weights reflects the functional similarities between the temporal gene expression patterns of overlapping genes and the gene set where they belong. They can be considered as “empirical gene set membership” specific to the given experimental condition. Lastly, we will develop an extension of Mann-Whitney U test that incorporates both weights and inter-gene correlation to perform hypothesis testing for gene sets. Extension for other rank-based tests will also be considered. We plan to apply this novel procedure to several publicly available time-course transcriptomic data to study strain specific mechanism involved in immune cell and cytokine mediated response to influenza infections. Based on our experience in influenza infection and vaccination related research, we believe the proposed method is perfect for studying strain-specific, condition-dependent variation of immune response to influenza virus because it avoids gene overlapping, has the flexibility of alter gene-sets for each subject.
2015 Faculty Awardees
Emily Carmody, MD
Assistant Professor of Orthopaedic
Co-Investigators: Michael Zuscik and Christopher Ritchlin
Project: Assessment of Forteo as a Therapeutic to Treat Knee Osteoarthritis
Traditional treatment strategies for Osteoarthritis are palliative, with the focus on pain management and joint replacement. Development of disease-modifying agents that can rejuvenate cartilage is a great unmet need. Thus, development of an effective remittive treatment for Osteoarthritis is a vital public health initiative with potential for tremendous impact. Data mined from the NIH-sponsored OA Initiative revealed improved WOMAC knee function scores in arthritic subjects coincidentally prescribed Forteo to treat osteoporosis. These preclinical and human data provide compelling rationale to study Forteo as a novel OA therapy directed at improving joint structure and function. The central Aim of this research program is to challenge the paradigm that cartilage loss in Osteoarthrtitis is irreversible.
David Herrmann, MD
Professor of Neurology
Project: A Pilot Study of Mexiletine for Muscle Cramps in Charcot Marie tooth Disease
Charcot Marie Tooth Disease (CMT) is a family of inherited peripheral neuropathies which affects 1/ 2500 individuals. CMT Type 1A (CMT1A) is an autosomal dominant disorder that accounts for 50% of CMT and manifests in childhood or early adulthood with progressive muscle weakness and atrophy, sensory loss, impaired ambulation, pain and disability. Muscle cramps affect about 85% of adults with CMT1A and impact quality of life and have been identified as an important therapeutic target in CMT1A. Mexiletine is an oral sodium channel blocker that in low doses has shown promise for prevention of muscle cramps, but data is lacking on its effectiveness in CMT1A. The overall goal of this pilot, double-blind randomized placebo controlled crossover study, is to obtain preliminary data on the efficacy and tolerability of low dose mexiletine for muscle cramps in adults with CMT1A.
Eva Pressman, MD
Chair and Professor of Obstetrics and Gynecology
Co-Investigator: Kimberly O'Brien
Project: Vitamin D Kinetics During Pregnancy
Nearly 30% of US women are either vitamin D insufficient or deficient. Vitamin D inadequacy during gestation is increasingly linked to adverse birth outcomes including preterm birth, risk of cesarean section and placental and pregnancy associated infections. At this time the IOM has not advocated any increase in vitamin D intake across gestation but this remains controversial in large part due to insufficient information on the basic physiology of vitamin D. Recent mass spectrometric instrumentation advances have provided opportunities to use deuterated vitamin D analogs as tracers to gain novel data on in vivo vitamin D metabolism at key life stages. In this pilot study, the overall objective is to take advantage of UHPLC-MS/MS instrumentation and deuterated vitamin D to obtain information on the absorption and half-life of vitamin D3 in non-pregnant and pregnant women.
Xingping Zhang, MD, PhD
Associate Professor of Orthopaedics
Co-Investigator: Stephen Kates
Project: Identification of the Effective Vascular Progenitors for Bone Repair and Regeneration
Stem/progenitor cell-based therapy has taken the center stage of regenerative medicine in the past two decades. A new cell-based therapy is emerging that aims to utilize endothelial progenitor cells (EPCs) alone or in combination with mesenchymal stem cells (MSCs) to enhance revascularization of the implant and thereby the survival and differentiation of the osteoprogenitors. Studies from several laboratories have demonstrated that delivery of EPCs alone or in combination enhances the vascularization of the implant and even contributes to the formation of bone in repair and reconstruction. However, despite the accumulating reports, the mechanisms by which EPCs participate in repair and the effective sources/populations of the EPCs that synergize with skeletal progenitors to enhance repair in vivo remain controversial and poorly defined. The goal of this pilot project is to devise a translational strategy to enhance bone repair and reconstruction.
2015 Clinical Trials Methods and Technologies
Martin Zand, MD, PhD
Professor of Medicine (Nephrology)
Co-Investigators: Jiong Wang, PhD, and John Treanor, MD
Project: Assessing Heterosubtypic Antibody Responses in Influenza Vaccine Clinical Trials
Pandemic influenza from emerging or mutated influenza strains is a large public health threat, and each year multiple clinical trials are done to assess vaccine efficacy. Current flu vaccination strategies confer strain specific immunity by inducing antibodies directed at the viral hemagglutinin protein, thus preventing virus binding to target cells and infection. Unfortunately, current methods of assessing vaccine efficacy are based on the hemagglutinin inhibition assay (HAI), which is 80 years old, time and labor intensive, and does not provide a continuous quantitative readout. Thus, there is a great need for an easy, rapid, sensitive and accurate assay to evaluate influenza vaccine efficacy, especially the induction cross-reactive immunity to multiple influenza strain subtypes (heterosubtypic immunity). The long-term goal of this project is to use this validation data to examine and track population heterosubtypic immunity, and to seek involvement in clinical trails of new H5 and H7 avian influenza vaccines..
2015 Trainee Awardees
Graduate Student in the Department of Medicine (Allergy, Immunology, and Rheumatology)
Mentors: Jennifer Anolik, Minsoo Kim, and Jane Liesveld
Project: Neutrophils as a driver of inflammation in lupus bone marrow
Systemic lupus erythematosus (SLE) is a debilitating autoimmune disease with a complex pathogenesis that presents a challenge for development of specific and effective therapeutic targets. This pilot project examines the role of one central mediator of SLE pathogenesis: chronically elevated type I interferon (IFN), examining both the mechanisms underlying its generation as well as its contribution to pathology in SLE marrow. The results of this pilot grant will lay the groundwork for development of specific therapeutic targets needed to treat pathology in lupus bone marrow.
2014 Faculty Awardees
Roman Eliseev, MD
Assistant Professor, Center for Musculoskeletal Research
Improving Mitochondrial Function in Mesenchymal Stem Cells to Accelerate Fracture Repair in Aging
The goal of this project is to test whether improving mitochondrial function in mesenchymal stem cells (MSC) will accelerate fracture healing during aging. In aging, MSC function and osteogenicity are compromised which is suggested to be a reason for delayed fracture healing. Our data and the literature indicate that MSC ability to differentiate into osteogenic lineage depends on their ability to activate mitochondria which are initially inactive in undifferentiated MSCs. Mitochondria in aged MSCs are less active due possibly to the MPT, a non-specific mitochondrial pore regulated by cyclophilin D and frequently observed in aged mitochondria. Thus, inhibition of the MPT is hypothesized to improve MSC mitochondrial function, osteogenicity, and, as a consequence, outcomes of fracture repair in aged mice.
Elizabaeth Guancial, MD
Assistant Professor of Medicine (Hematology and Oncology)
Chemoprevention of bladder cancer through estrogen receptor modulation
While bladder cancer (BC) has not historically been viewed as a hormone-sensitive cancer, differences in rates of development and prognosis between men and women with BC suggest that estrogens or the estrogen receptor (ER) may be involved in BC carcinogenesis. In vitro studies in BC cell lines demonstrate ER-dependent growth inhibition by antiestrogen agents. Most patients with muscle-invasive BC are unable to receive recommended neoadjuvant or adjuvant chemotherapy due to medical comorbidities and toxicity, despite a high risk of relapse after radical cystectomy alone. Therefore, new treatments are urgently needed to reduce the risk of BC relapse after surgery and for the treatment of advanced BC in patients with other comorbidities. Antiestrogens are commonly used to treat breast cancer and have an acceptable safety profile for most patients. The objective of this project is to investigate the therapeutic role of antiestrogens in the chemoprevention and treatment of BC in order to identify novel therapies that are effective and tolerable and to establish a mechanism of action for these agents through the study of the relative contribution of the two ER subtypes, ERalpha (ERa) and ERbeta (ERb), to BC carcinogenesis in order to identify predictive biomarkers of response to antiestrogens.
R. John Looney. MD
Professor of Medicine (Allergy, Immunology and Rheumatology)
Role of the Gut Microbiome in Preventing Allergic Disease
The epidemic of allergic and autoimmune diseases in developing countries is one of the greatest medical challenges of the 21st century. Although we have greatly improved treatment for many of these diseases, our goal should also be prevention. As discussed above there is considerable data suggesting that the key to this epidemic his how the environment influences immune system development. Finding a population at low risk and comparing immune system development in that population to immune system development in a high risk population is a critical need for this entire area of investigation. The Old Order Mennonite's (OOMs) population of upstate New York provides and an ideal low risk population. The OOMs have a lifestyle incorporating all the various environmental factors that have been associated with a low risk of asthma and allergic diseases including growing up on a farm, having large families, exposure to numerous pets in farm animals, exposure to raw milk, low rate of smoking, and low rate of antibiotic utilization. Our preliminary studies have confirmed at the OOMs did have a markedly lower risk of asthma in the general population in upstate New York. Central Hypothesis – The low rate of atopic disease in children who grow up on farms with numerous siblings is related to accelerated maturation of the immune system due to stimulation of the mucosal immune system early in life by the a diverse microbiome that stimulates innate immune receptors.
Edward Messing, MD
Professor of Urology
Exosomes from bladder cancer patients can serve as biomarkers of disease progression
Bladder cancer is the 5th most commonly diagnosed cancer, the most expensive to treat over the lifetime of the patient, and utilizes the most Medicare dollars. Much of the cost associated with bladder cancer is related to the surgical interventions necessary to diagnose and treat the disease. Moreover, treatment of high-grade bladder cancer is marked with elevated rates of morbidity and mortality (i.e. 36% 5 year survival for pT2 disease). Over the last thirty years there has been very little advancement in chemotherapeutic options for bladder cancer. Identifying markers of tumor progression through less invasive means could expedite treatment, prevent progression, identify novel therapeutic targets and contain cost. Recently an interest in small membrane bound vesicles called exosomes has emerged. Exosomes have been shown to be important mediators of tumor progression and contain biologically active proteins, messenger (m)RNA, long non coding (lnc)RNA, and micro (mi)RNA. Importantly, exosomes can be readily isolated from blood and urine. We have identified lncRNA and mRNA associated with tumor progression in the exosomes of patients with pT2 bladder cancer suggesting the feasibility of this project. The fundamental goals of this project are to identify stage-specific biomarkers of bladder cancer progression by RNA-sequencing of primary tumors as well as exosomes purified from the urine and blood of patients, and ultimately in downstream experiments identify which of these RNA are important in tumor progression and therefore may serve as targets for anti-sense therapeutics.
Craig Morrell, DVM, PhD
Associate Professor of Medicine
Novel microRNA Based Therapy to Improve CD4+ T-cell Responses to Vaccination
This project will explore how miR-451 regulates T-cell responses to malaria infection and the use of antagomirs to increase responses to malaria vaccination or infection, representing an important conceptual and therapeutic advancement. These studies will be catalytic in generating new programs and funding for our clinically applicable research. An additional goal of our program is the stimulation of continued crossdisciplinary collaborations between members of the CVRI and Microbiology and Immunology. We will use the combined expertise and knowledge of the Morrell lab, who have extensive experience in animal models of malaria infection), the Lowenstein lab who have published many studies related to miRNA, and the Fowell lab who have great expertise in mechanisms of CD4+ T cell responses.
Sherry Spinelli, PhD
Research Associate Professor of Pathology and Laboratory Medicine
The Role of Microparticle-Derived Thy-1 (CD90) in Type 2 Diabetes Mellitus
Thy1 (CD90) is a glycophosphatidylinositol-anchored protein that was discovered decades ago, and recognized simply as a surface marker of unknown function. Recently, our laboratory pioneered studies demonstrating that Thy1 is a key signaling protein that inhibits adipogenesis (fat formation). Thy1 expression down regulates crucial pro-adipogenic factors, such as peroxisome proliferator activated receptor gamma (PPARgamma). While Thy1 was originally identified on the surface of nucleated cells, we have discovered it is present on anucleate platelets and on the platelet progenitor cell, the megakaryocyte. Importantly, Thy1 is also released in platelet microparticles (MPs), thus Thy1 could control adipogenic potential via transcellular regulation in recipient cells. Given the importance of Thy1 in the regulation of adipogenesis and attenuation of proinflammatory adipokines, Thy1 may be an integral player in the pathophysiology of type 2 diabetes mellitus (T2DM), an emerging global epidemic characterized by obesity and a proinflammatory profile. Our group recently detected and measured Thy1 expression in megakaryocytes, platelets and MPs in type 2 diabetics (T2D) versus healthy individuals. Levels of Thy1 were much lower in T2Ds, and importantly, the lack of Thy1 in T2D MPs could be a crucial mediator in upregulating inflammation and adipogenesis in recipient cells.
2014 Trainee Awardees
Specialized proresolving mediators act as novel therapeutics against infection
Nontypeable Haemophilus influenzae (NTHi) is a gram-negative, opportunistic pathogen that commonly causes respiratory diseases, including bronchitis and pneumonia. People with a preexisting inflammatory condition, such as chronic obstructive pulmonary disease (COPD) or an additional infection, are particularly susceptible to NTHi. These infections are increasing in incidence and are often persistent, resulting in bacteria propagating in the airways. Recently, endogenously produced, specialized proresolving lipid mediators (SPMs) were discovered. SPMs play a critical role in the active resolution of inflammation through both anti-inflammatory and pro-resolving actions and are thus strong candidates for use in treating infections. They have been shown to be efficacious in reducing mortality and in decreasing bacteria blood levels through enhanced phagocytosis in mice infected with E. coli. The efficacy of SPMs in promoting resolution of pulmonary infections, however, has not been investigated. My unpublished data shows that SPMs can dampen lung inflammation caused by cigarette smoke or LPS in preclinical mouse models. In human macrophages SPMs increase phagocytosis of bacteria and apoptotic neutrophils.. This research is the first to assess SPMs in pulmonary infections and will provide the groundwork for further investigation and eventual translation of SPMs into a clinical setting.
Comparative effectiveness of screening methods for type 2 diabetes: a pilot study
The ultimate goal of this project is to compare, in a large-scale randomized clinical trial, the effectiveness of three screening strategies to detect type 2 diabetes: fasting plasma glucose (FPG), hemoglobin A1c (HbA1c), and point of care testing (POCT) HbA1c, as they are performed in real clinical settings. Currently, more than one quarter of adult diabetics in the US are undiagnosed and thus at risk of developing diabetes complications by the time of their delayed diagnosis. The panel of approved tests for screening and diagnosis of type 2 diabetes includes only tests performed at central laboratories, such as FPG and HbA1c. Although not approved for this purpose, POCT HbA1c, which is performed at the practice’s site and provides the results within minutes of testing, is used to screen for type 2 diabetes in some clinical settings. Previous studies have focused on the laboratory performance of the screening tests for diabetes, but have not compared their effectiveness when used in real clinical settings.
2014 NBEM Awardees
Associate Professor of Biostatistics and Computational Biology
Cure Model in Recurrent Event Data
Although cure model has been studied in traditional survival analysis, for recurrent event data, the case is much more complicated. Unlike the survival data with at most a single event, the observations are censored for study subjects. Some subjects may be cured from the very beginning of the study while some subjects are cured after a few relapses. In this study we will (1) develop a method to test whether the follow-up is long enough to declare some individuals cured; (2) develop nonparametric method to estimate the cure fraction; (3) develop semiparametric method to study the effects of come covariates on the cure probability; and (4) develop a method to make prediction of cure based on some prognostic factors.
Assistant Professor of Biostatistics and Computational Biology
Causal Inference with Zero-inflated Predictors: Alcohol, HIV Risk, and Depression
To make causal inference, we have to compare outcomes between the at- and non-risk groups
separately derived from the theoretical concept of structural zeros. The challenge is that the membership of the two risk groups is unknown for subjects without observed drinking (e.g., 0 in the days of drinking); they may or may not be at-risk, depending on whether they have structural or random zeros. Some recent studies modeled the zero-inflated nature of the count response when they appear as the dependent (or response) variable, but no statistical methodology is available to infer the causal relationships when such variable appears as independent variable. Our goals in this proposal are (1) develop new statistical methods to model the latent membership of the at and non-risk group to enable causal inference of the risk (i.e., an independent variable) on health outcomes; (2) create software for implementing the new statistical methods; (3) test the hypotheses about the role of alcohol on causing (a) depression, and (b) greater frequency of HIV risk behaviors using HANES 2009-2010 data; (4) document the findings as pilot results for our R01 resubmission.
Research Assistant Professor of Biostatistics and Computational Biology
Modeling Human Interactions in Social Networks
The proposed research addresses the lack of methods for modeling and assessing the effect of human interaction on behavioral and health outcomes of socially connected individuals. A main premise of this pilot study draws upon recent innovations and methods in the evolving field of data science, a new and burgeoning field that integrates both qualitative and quantitative data analytics to draw and capitalize upon Big Data possibilities stemming from mobile and web technologies and social media. We stand on the cusp of a new opportunity due to: improvements in computing, the rapidly growing ubiquity of smart environments, and a resulting duality in contemporary society, that is, the norm of living simultaneously within virtual as well as geophysical proximal interactions and community contexts. Massive amounts of diverse unstructured and structured data about interactions and interventions can now be culled from online sources such as social media, emails, and websites. Mobile devices and online social media such as Smart phones and Twitters generate new potentials over the traditional research paradigm for virtually all disciplines including statistics and its applications to clinical research and practice. Traditional statistical paradigms premised upon individual, within-subject attributes are fundamentally at odds with dependent human interconnections in such data. Well-developed theories of U-statistics and Functional Response Models, along with our experience with leading these theories and applications, positions the proposed proof of concept study to pilot test a new paradigm to model the effect of human interaction on changes of behavioral and health outcomes of socially connected individuals. Further, the software and associated documentation, to be made available via a number of venues including "CTSpedia.org" (a new NIH-funded reference and resource website with a repository of statistical functions to promote multidisciplinary interactions and collaborations), will catalyze the use of such interpersonal, participatory and interactive information from online and mobile sources in disease prevention and health promotion.
2013 Faculty Awardees
W. Richard Burack, MD, PhD
Associate Professor, Pathology and Laboratory Medicine
Quantifying Tumor Diversity to predict and target Cancer progression
Biomarkers that predict chemotherapy resistance, relapse and/or transformation risk are critically needed. Because intratumoral genetic diversity is the basis for the evolution of chemotherapy
resistance, relapse, and aggressive transformations of cancers, a metric of genetic diversity has the potential to be a transformative biomarker. Genetic diversity has long been recognized in Follicular lymphoma (FL), a B cell non-Hodgkin Lymphoma that is generally indolent but has a propensity to suddenly transform into a highly malignant form. Follicular B cells normally express an intrinsic genome-damaging enzyme, the APOBEC family member Activation-Induced (Cytidine) Deaminase (AID), an activity required for immunoglobulin diversification. While mutations associated with cancer frequently have the signature sequence of APOBEC/AID targeted loci, data directly implicating AID-induced damage in the generation of intra-tumoral diversification are lacking. Demonstrating this association would suggest a powerful biomarker for risk of progression: greater diversity predicting a greater risk of progression. Making this link requires a quantitative assay that measures AID-dependent intra-tumoral genetic diversity in a high throughput fashion from patients’ specimens, and such an assay has not been reported. Dr. Burack has developed a novel, DNA sequencing based method and analytical approaches to quantify intra-tumoral genetic diversity attributable
to AID. Dr. Burack’s group will apply this method to typical human tumor specimens to directly test if diversity predicts cancer progression.
Laura Calvi, MD
Associate Professor, Medicine (Endocrinology)
Osteoblastic Function in Human Leukemia
The mechanisms by which a leukemic clone suppresses normal hematopoiesis are poorly understood, and yet this phenomenon likely contributes to disease progression, disease morbidity and response to therapy. A recent analysis of the bone marrow microenvironment (BME) in a syngeneic mouse model of acute myeloid leukemia demonstrated dramatic osteoblastic defects. Dr. Calvi's laboratory has demonstrated the central role of osteoblastic lineage cells in hematopoietic stem cell (HSC) regulation, these data identify osteoblastic cells as a potential clinical target to stimulate normal HSC recovery in leukemia and decrease BME support of leukemic stem cells (LSCs). Moreover, they discovered leukemic production of the chemokine CCL3, which inhibits osteoblastic function in multiple myeloma. The goal of this pilot project is improving normal hematopoiesis and decreasing microenvironmental support for Leukemic Stem Cells, efficiently, effectively and safely apply pharmacologic tools currently approved for bone anabolic
treatment to leukemia. Data from this project would represent a paradigm shift in the therapy for patients with AML, where targeting of the BME improves our ability to treat the leukemia and more readily restore normal hematopoiesis.
Alan Smrcka, PhD
Professor, Pharmacology and Physiology
Inhibition of G protein beta/gamma signaling as a therapeutic approach to treatment of lupus
In complex autoimmune diseases such as Systemic lupus erythematosus (SLE) or rheumatoid arthritis the pathologies are driven in part by alterations of many circulating factors and responsiveness of cells to these factors. Inhibition of a shared signaling mechanism downstream from these receptors, that operates both in the adaptive and innate immune system, will likely result in higher efficacy than specific pathway or factor targeting. One such pathway is the G protein beta/gamma subunit signaling pathway downstream of the chemokines receptors that control the migration activation and survival of all types of immune cells. Dr. Smrcka has identified a compound that inhibits G protein beta/gamma subunit-dependent signaling in isolated human neutrophils, inhibits neutrophil migration in vitro, and inhibits acute inflammation in mice by preventing neutrophil migration and activation at sites of inflammation. Dr. Smrcka plans to test the viability of Gbeta/gamma inhibition as a treatment paradigm for lupus and to test a novel hypothesis that Gbeta/gamma inhibition ameliorates disease by acting at both the innate and adaptive immune system. Dr. Smrcka will collaborate with Jennifer Anolik, MD, PhD, Associate Professor of Medicine (Allergy/Immunology and Rheumatology) on this project.
2013 Trainee Awardees
Hsi-min (Jim) Hsiao, BS, MS
Pathology and Laboratory Medicine
Novel pro-resolving lipid mediators reduce cigarette smoke-induced emphysema
Chronic obstructive pulmonary disease (COPD, emphysema and chronic bronchitis) is the fourth leading cause of death in the United States. Importantly, the disease continues to worsen even after smoking cessation. Current therapies for COPD attempt to relieve the symptoms but do not alter the course of the disease; therefore, new therapies for COPD are desperately needed. This study will provide critical pre-clinical data needed to prepare for human clinical trials of resolvins in lung disease. A multidisciplinary collaboration has also been established to analyze the effects of resolvins on lung function, inflammatory response, cell death and signaling cascades, including clinically relevant measures such as pulmonary function testing, to increase the translational potential.
Jonathan Stone, BA, MD
Intraparenchymal Stent for Obstructive Hydrocephalus (IPSOH): a Novel Technology
Hydrocephalus is a common debilitating neurologic disease affecting a significant portion of the pediatric and adult population. The current surgical treatment options are fraught with complications and excessive costs heralding the need for new technology. Furthermore, the pathophysiological effects of hydrocephalus and fluid shunting on brain interstial fluid are unknown and need further investigation to improve patient care. This project will not only test the efficacy of this new shunt system in an animal model, but will also evaluate the movement of interstial fluid in hydrocephlaus and after both interventions.
2013 UNYTE Awardees
Steven Bernstein, MD
Professor of Medicine (Hematology and Oncology)
Lymphoma and its microenvironment; a novel in vivo model to study its interplay
Follicular lymphoma is an incurable disease with conventional therapy and thus new approaches for treatment are needed. As the lymphoma cells require signals from the other non-malignant cells in the tumor (the tumor microenvironment) to survive, targeting such interactions represents a novel approach for treatment. Recent data shows that the FL immune microenvironment, particularly the distinct T-cell populations infiltrating the tumor, play a critical role in modulating the biology and clinical behavior of this disease; however an understanding of how these populations modulate FL B-cell growth, viability and sensitivity to immune-chemotherapy (IC) is lacking. The investigators are now poised for the first time to test the central hypothesis that the interplay of FL Tregs and Tfh either directly or indirectly modulate FL B-cell growth, viability and sensitivity to IC in vivo.
Ankur Chandra, MD
Assistant Professor of Surgery (Vascular Surgery)
Regional Ultrasound Wall Strain Measurements to Predict Risk of AAA Rupture
Two-hundred thousand new Abdominal Aortic Aneurysm (AAA) cases are diagnosed each year in the United States; fifteen thousand people die from AAA rupture each year, making it the 13th leading cause of death in this country and affecting 1 in 250 individuals over 50 years of age. The exact cause of AAA formation is still unknown, although many theories base their pathogenesis as a multifactorial cause. Creating a “strain fingerprint” to determine the probability of a rupture is a viable new option that could significantly decrease AAA deaths. The project goal is to develop a novel application of existing ultrasound strain algorithms as a transcutaneous imaging modality to predict the risk of AAA rupture, regardless of size.
2013 NBEM Awardees
Anthony Almudevar, Bsc, Msc, PhD
Associate Professor of Biostatistics and Computational Biology
Predictive Models for Longitudinal Technological Home Monitoring Data
The aim of this proposal is to develop preliminary data and a proof-of-concept demonstration to leverage future research. The CB assessment application is particularly suitable for a number of reasons. The number of alternative assessment tools is limited to self-reporting, psychometric testing, or direct interview. We note also the availability of processed data from two parallel monitoring systems for caregiver/patient dyads, which is a highly specialized and uncommon scenario.
Changyong Feng, PhD
Associate Professor of Biostatistics and Computational Biology
Allowance for center effects in the analysis of randomized clinical trial with time-to-event outcomes
Many randomized clinical trials (RCT) have time-to-event outcomes. The log-rank test is widely used to analyze such event-time data as it is the most efficient nonparametric test under the hypothesis of proportional hazards. However the log-rank test assumes that individuals in the same treatment group are all homogeneous. Heterogeneity among individuals in a randomized study does not invalidate the log-rank tests, but it may make it less efficient. It is common to control heterogeneity using a stratified log-rank test (SLRT). It is known that if there is substantial heterogeneity among centers, the SLRT will be more sensitive to treatment differences than the unstratified test (here denoted ULRT). On the other hand, unnecessary stratification can lead to a loss of efficiency. However the trade-off between these two situations is still not well understood. In some practical situations the ULRT appears to be more sensitive than the SLRT even when there is quite substantial heterogeneity between Centers. In this proposal we will compare the relative efficiency of SLRT and ULRT under two different scenarios and obtain an optimal linear combination of these test statistics which maximize the power.
Xing Qiu, PhD
Assistant Professor of Biostatistics and Computational Biology
A Unified Method for Differential Expression and Differential Association Analyses
Thousands of basic research projects use the microarray technology, yet very few of them have been successfully translated into clinical applications. This proposal responds to this challenge by integrating normalization, DE analysis, and DA analysis, in such a way that not only the computational cost is reduced, but also the false positives/negatives are reduced by using one MTP for both analyses simultaneously.
2012 Faculty Awardees
Neil Blumberg, MD
Professor, Pathology and Laboratory Medicine
Improving Platelet Storage and Transfusion Outcomes with PPARγ Ligands
Platelet transfusion is the most commonly used therapy for patients with trauma, hematologic diseases or cancer who are experiencing bleeding and low platelet counts. Our proposed investigations are vitally important, since in the USA alone, almost two million platelet transfusions are given each year. Poorly understood mechanisms that occur during platelet storage, termed the “platelet storage lesion”, reduce platelet transfusion efficacy and safety. Consequently, patients are transfused not only with partially or abnormally activated platelets that reduce transfusion efficacy, but also with storage supernatants containing many potentially harmful bioactive mediators that can elicit adverse responses to transfusion. For example, platelet transfusion can cause alloimmunization, fever, rigors, and allergic reactions. A critical barrier to the prevention of adverse post-platelet transfusion events is the absence of approaches to modify storage conditions that ameliorate the platelet storage lesion. Hence, there is an urgent need to investigate new targets to attenuate platelet activation mechanisms, thus improving the efficacy and safety of platelet storage.
Our objective is to focus on novel aspects of platelet biology to better understand the mechanisms that drive unwanted platelet activation during storage, leading to the development of new technologies of platelet storage that will maintain normal platelet quality and function.
Lisa DeLouise, PhD, MPD
Associate Professor, Dermatology
High throughput sorting of rare cells from blood using Microbubble Arrays
Development of monoclonal antibodies (mAbs) for therapeutic use is a rapidly growing $50 billion/year market. Hybridoma technology is a time tested technique used to generate antibodies; but it is costly to screen all clones generated and therefore quality antibodies
maybe missed. Non-animal techniques that can identify and characterize ASC in peripheral human blood that exhibit high binding specificity and affinity are in high demand. Microbubbles are novel compartments formed in an optically clear elastomeric material. MBs exhibit unique properties for cell culture which are leveraged in this screening application that provides many advantages over existing techniques.
Our project seeks to advance development of a new high throughput cell screening technology based on microbubble (MB) arrays. Based on preliminary studies we will investigate the development of MB arrays to enrich, identify, characterize, and recover rare antigen specific antibody secreting (ASC) from peripheral blood. This project will investigate the limits of this assay and to automate the data acquisition and analysis.
Ajit Kulkarni, PhD
Research Assistant Professor, Medicine (Pulmonary/Critical Care Division)
Characterization of antifibrotic effects of CDDO, a small electrophilic compound
Pulmonary Fibrosis compromises normal lung function and structure due to scarring of lung tissues. Scarring is caused by proliferation of fibroblasts and myofibroblasts, and excess deposition of extracellular matrix proteins in fibrotic foci. There is an urgent unmet need to develop new therapies for pulmonary fibrosis since effective treatments are often lacking.
We have reported that a small electrophilic compound, 2-cyano-3,12-dioxoolean-1,9-dien-28-oic acid (CDDO) inhibited the transforming growth factor (TGF)-β induced differentiation of human lung fibroblasts to myofibroblasts (scar forming cells) in vitro. We hypothesize that by inhibiting myofibroblast differentiation and proliferation, and by inhibiting expression of pro-fibrotic genes by fibroblasts and myofibroblasts, we will be able to slow or arrest the progress of the disease in patients with lung fibrosis. Here, we will investigate the efficacy of CDDO in vivo using models of pulmonary fibrosis. We hope these studies will rapidly lead to a new therapy for patients with suffer from pulmonary fibrosis.
Yang Liu, BM, PhD
Research Assistant Professor, Neurosurgery
Therapeutic targeting of CXCR7 in malignant glioma by small molecule antagonist
Malignant gliomas represent a uniformly fatal form of cancer. Despite advances in neurosurgical techniques, chemotherapeutic regimens and radiotherapy protocols, little improvement has been made in the 5-year relative survival rates of brain tumor patients during the past several decades. Glioblastomas, the most common and highest grade of malignant glioma, are highly vascular, highlighting a potential therapeutic target. Chemokines and their receptors play critical roles in many physiological and pathological processes, including brain cancer. Chemokine receptor 7 (CXCR7) was recently identified as second receptor for stromal cell derived factor 1 (SDF1) and exerts an important role in tumor growth and vascularization.
Previously, we found that CXCR7 mRNA was expressed at levels 9 times higher in brain tumors than normal brain samples and was localized to vascular regions within glioma samples. We also found that inhibition of CXCR7 expression by targeted siRNA significantly impeded glioma cell proliferation and motility in vitro and limited intracranial xenograft growth and improved mouse survival, validating CXCR7 as a potential therapeutic target for glioma. Furthermore, recent studies have shown that SDF1 can recruit bone marrow derived endothelial progenitor cells to tumor neovessels and attract haematopoietic progenitor cells to intracerebral glioma. Therefore, we propose that SDF1/CXCR7 play an important role in brain tumor growth and maintenance of the vascular niche between brain tumor stem cells and the neurovasculature.
2012 Trainee Awardee
PhD Candidate (Epidemiology), Public Health Sciences
Developing an age-specific decision scheme for prehospital triage of injured older adult
Injury is among the leading causes of death and disability for older adults. Treatment at advanced care hospitals specializing in injury, also known as trauma centers, has been shown to significantly improve patient outcomes. However, the selection of a receiving hospital is dependent upon emergency medical services (EMS) providers making appropriate clinical judgments in the prehospital setting – a process referred to as trauma triage. It is known that older adults are less likely to receive trauma center care than younger adults, but reasons for this age-based disparity are not well understood. Evidence suggests two mechanisms are involved: 1) EMS providers' decision-making process differs for older adults compared to younger adults; and 2) the current guidelines to aid EMS providers in their trauma triage decisions are inadequate to identify older adults who require trauma center care. This proposal aims to assess both of these potential reasons by using a combination of analytic methods. As the number of older adults in the US is projected to increase dramatically in future years, injury will continue to be a major burden on the public's health. Identifying and addressing reasons for the age-based disparity in trauma center care is vital to improving patient outcomes for this population.
2012 UNYTE Awardee
Katia Noyes, PhD, MPH
Professor, Public Health Sciences
Validity of Self-Reported Data for Studying Cognitive Problems and Depression
MS is the most common neurologic disease affecting young adults, striking nearly 500,000 people in the US. MS-related symptoms include physical disability, fatigue, cognitive impairment, and affective disorders. MS is different from most of other chronic conditions: its financial impact associated with person's productivity, social functioning, and employment is nearly as significant as the economic burden of medical treatment. The prevalence of major depression in patients with MS (16%) is over twice that among chronically ill population without MS (9%), and nearly four times higher than in general population (4%). Hence, it is critical that the research community takes concrete steps toward resolving the uncertainty surrounding the optimal treatment of individuals suffering from MS, particularly those with affective and cognitive dysfunction.
The main goal of this study is to assess validity of self-reported information about cognitive and mental health status in patients with multiple sclerosis (MS) and to understand feasibility of using these data for quality of care assessment and program evaluation. This study aims to stimulate and enhance our cross-disciplinary collaboration between the Departments of Public Health Sciences and Neurology, University of Rochester and Baird MS Center in Buffalo, NY. The study will involve two sites (University of Rochester and University of Buffalo/Jacobs Neurologic Institute in Buffalo, NY) of the New York State Multiple Sclerosis Consortium, one of the largest databases of MS patients.
2012 Novel Biostatistical Epidemiological Methods (NBEM) Awardees
Hua He, PhD
Assistant Professor, Biostatistics and Computational Biology
Novel models for analyzing drinking outcomes: A pilot study comparing competing approaches
The COMBINE Study was conducted from 2001 to 2004, with 1,383 individuals of alcohol dependence assigned to one of nine pharmacological and/or psychosocial treatment conditions. The only other alcohol treatment study in the U.S. on this scale was Project MATCH, conducted in the early 1990’s. COMBINE compared two promising pharmacological treatments for alcoholism, naltrexone and acamprosate, alone and in combination with an combined behavioral intervention (CBI). Although the primary outcome papers of COMBINE have been published, researchers are at an early stage in exploiting the potential of this dataset to address questions beyond a comparison of its treatment conditions.
This application addresses an important statistical issue in alcohol research: the analysis of a bounded count response with structural zeros and overdispersion within a longitudinal data setting. This issue is highly relevant to the analysis of treatment effectiveness of drinking interventions for various risk populations. We will develop a new approach for such zero-inflated binomial-based (ZIB) count response for cross-sectional and longitudinal studies, and use it as a benchmark to evaluate the performance of the approaches that either have been used for analyzing such count responses in the alcohol research literature or existing alternatives in the statistical literature, such as zero-inflated Poisson (ZIP) model, general regression model for transformed count response and Hall & Zhang's marginal models for ZIB, by conducting intensive simulation studies and applying them to drinking outcomes in COMBINE.
Rui Hu, PhD
Research Assistant Professor, Biostatistics and Computational Biology
Detecting Intergene Association Changes in Microarray Data
Microarray technology has become a routine gene expression analysis tool in recent years. Biomedical researchers rely on this technology to identify potentially “interesting” genes. Typically, individual genes are tested for their differential expressions between phenotypes by the two-sample Student’s t-test or its nonparametric counterpart. The resulting p-values are adjusted by a chosen multiple testing procedure (MTP) in order to control certain group-wise Type I errors.
We plan to develop a novel gene selection procedure based on intergene association structure changes across different phenotypes. Gene differential association analysis which was explored in our preliminary study utilized the gene association vector in the gene selection, which provided quite conservative testing power. In this study, we will focus on gene pairs and search for most powerful statistical tests to detect differential associated genes.
Yinglin Xia, PhD, MS
Research Assistant Professor, Biostatistics and Computational Biology
Integrative Analysis of Pathways to SA and PPD in High Risk Families
This proposal is to develop a new class of statistical models to facilitate integrative analysis of multi-faceted data and to illustrate the new methodology by applying it to examine pathways to suicide attempts (SA) in high risk families using the GenRED database. The feature-rich GenRED database provides a rare and unique opportunity to explore the types of risk factors and what roles they play in the pathways to SA. Significant advances have been made over the past few decades in the theory and applications as well as software development for fitting structural equation models (SEM). However, our recent work shows that there are several limitations in existing methods.
Our goals in this proposal are to (1) test hypotheses of SA using existing standard SEM, (2) develop a class of distribution-free SEM, (3) test hypotheses using the new methods. Specifically, we will create a dataset using GenRED to examine the following set of hypotheses concerning pathways to SA using existing SEM.
2011 Faculty Awardees
Nancy Bennett, MD, MS
Director of the Center for Community Health
Co-Investigators: Jennifer Carroll, MD, MPH, Linda Clark, MD, MS, Michael Nazar, MD, and James Sutton, PA.
Project: Comparative Effectiveness of practice-based diabetes prevention programs
The Healthy Living Program, a community-developed program, adapted from the evidence-based Stanford program, and conducted in community settings has been ongoing in the Rochester community since 2001 and has had over 1,700 participants in both the original program targeting African Americans and the revised program, Vida en Salud, for the Latino community. We propose to compare the effectiveness and costs of the two programs through a randomized trial conducted at primary care practices. In addition, we will study the feasibility of collecting data regarding the behavioral/motivational mechanisms through which these programs are successful. This pilot will enable the team to design the optimal study, refine endpoints and measurement instruments, explore the feasibility of randomization in a community health center clinical setting, and collect pilot data to determine effect sizes.
John Frelinger, PhD
Professor of Microbiology and Immunology
Co-Investigator: Mark Sullivan, PhD
Project: Protease activated cytokines: a novel methodology for the delivery and activation of cytokines
Cytokines play critical roles in cellular immune responses. The immunotherapy of cancers with cytokines has had some dramatic clinical successes, but side effects limit their use when delivered systemically. We are developing a novel approach that employs a fusion protein (FP) in which a cytokine is joined to its specific binding moiety; an antibody fragment (scFv) identified using phage display, separated by a protease site. The strategy is that before cleavage, the cytokine is largely inactive, but that after cleavage by a protease expressed at the tumor site, the cytokine can become available to interact with high affinity receptors on immune cells.
Sherry Spinelli, PhD
Research Associate Professor of Pathology and Laboratory Medicine
Co-Investigators: Richard Phipps, PhD, Charles Francis, MD, and Stephen Hammes, MD, PhD
Project: Microparticle miRNAs as transcellular messengers in diabetes and vascular disease
Microparticles (MPs) are submicron-sized membrane vesicles that are released into the blood by platelets and vascular cells. MPs contain cellular information in the form of bioactive proteins, lipids and molecules that influence cells, not only in the region of their release, but are carried in the circulation to elicit broad-reaching transcellular effects. A major knowledge gap is in understanding the mechanisms that govern MP transcellular communication. This research plan will investigate MPs derived from healthy and type-2 diabetic individuals. The hypothesis is that altered packaging of miRNAs in platelet MPs is a key element in vascular cell communication that may promote inflammation. We predict that differences in miRNA levels and types could serve as biomarkers of disease progression and lead to therapeutic strategies to modulate cellular dysregulation.
2011 Trainee Awardee
Roni Kobrosly, MPH
PhD Candidate in the Department of Public Health Sciences (Epidemiology)
Co-Investigators: Edwin van Wijngaarden, PhD, (primary mentor), Jan Moynihan, PhD, Deborah Cory-Slechta, PhD, Christopher Seplaki, PhD
Project: Examining the link between allostatic load and depressive symptoms among the elderly
Allostatic load is a developing epidemiologic concept that has been used to quantify the physiologic costs of cumulative life stress, whether psychological or physical. Although various forms of life stress have been linked with late-life depressive symptoms, the association between allostatic load and depressive symptoms has never been assessed. This proposal entails a cross-sectional study of approximately 200 community dwelling older adults examining the complex relationship of life stress, allostatic load, and psychosocial factors with the severity of late-life depressive symptoms. One specific aim and one exploratory aim are detailed: (1) to examine the relationship between allostatic load and the severity of depressive symptoms, and, as an exploratory aim, (2) to conduct a path analysis to examine the complex causal web of biological, psychological, and social factors underlying late-life depressive symptoms.
2011 UNYTE Awardees
Richard Burack, MD, PhD
Associate Professor of Pathology and Laboratory Medicine
Co-Investigator: Robert Hutchison, MD (SUNY Upstate Medical University)
Project: Regional stat tumor procurement to support studies of lymphoma
A major barrier to studying cancer and its treatment is the limited availability of human tumors. To address this problem, the NCI funds biorepositories such as the lymphoma-specific biorepository at URMC, a component of the NCI-funded SPORE in lymphoma at URMC. While tumor “banks” are common, the focus of the URMC biorepository on viable specimens is perhaps unique in the country, and has been critical to obtaining NCI-funding for projects at URMC. The availability of this resource has spurred interest in studying lymphoma, all with important technologies and/or hypotheses that could be reasonably tested given sufficient materials. However, current NCI-funded programs are using essentially all the specimens obtained at Strong Memorial Hospital/URMC. Demonstrating a regional lymphoma biorepository focused on stat distribution of living lymphomas to researchers will be critical to several proposals under development for submission to the NCI and NIH.
Janet Dehoff-Sparks, PhD
Professor of Pathology and Laboratory Medicine
Co-Investigator: Michael Greene, PhD (Bassett Research Institute)
Project: Protein kinase C (PKC) activation and inhibition of VLDL triglyceride (TG) export
Obesity is associated with non-alcoholic fatty liver disease (NAFLD) which can progress to the more serious condition of non-alcoholic steatohepatitis (NASH), a known precursor to cirrhosis and hepatocellular carcinoma in humans. In this proposal we examine the extent to which specific hepatic PKC isoforms regulate hepatic TG balance. Results will provide evidence to support a pharmacologic approach to reduce hepatic steatosis by blocking PKC signaling specifically in the liver thereby reducing lipogenesis and enhancing VLDL TG export.
2011 Novel Biostatistical Epidemiological Methods (NBEM) Awardees
Rui Hu, PhD
Research Assistant Professor of Biostatistics and Computational Biology
Co-Investgators: Sandhya Dwarkadas, PhD, Galina Glazko, PhD, Xing Qiu, PhD
Project: Clustering Differentially Associated Genes
We plan to develop a novel gene clustering procedure based on gene differential association analysis which was explored in our preliminary study. By applying this procedure to microarray gene expression data, we can search for differentially associated gene groups such that genes belonging to the same group do not change their association structure across different phenotypes while the association structure of genes from different groups are differentiated across phenotypes. This novel procedure has the ability to uncover biologically meaningful gene groups which contain differentially associated genes. Consequently, it can complement and enhance the existing clustering algorithms based on differentially expressed genes. It will also help us understand how the phenotypic differences of gene dependence structure can be used to cluster genes into biologically meaningful units.
Hongqi Xue, PhD
Research Assistant Professor of Biostatistics and Computational Biology
Project: Parameter estimation for nonlinear stochastic differential equation models from noisy longitudinal data in HIV dynamic research
In AIDS research, one major area is to model the interaction between HIV virus and the immune cellular responses. It is very useful for understanding the pathogenesis of HIV infection and assessing the potency of antiviral therapies. The investigator is focusing on the following four aims: (1) To develop multidimensional nonlinear SDE models for modeling the interaction between HIV virus and the CD4+ T cells for HIV-infected patients under treatment; (2) To develop parameter identifiability methods for such models with noisy longitudinal data; (3) To develop novel parameter estimation methods with sound theoretical justifications and computational efficiency for such models; and 4) To develop novel model selection methods between SDE models and ODE models.