Using statistics to fight infection?

NIH-NIAID supports development of mathematical models and statistical methods as important tools to fight against infectious diseases

IMS Bulletin Volume 40 . Issue 2, March 2011

Hulin Wu IMS Bulletion v40 i2The National Institute of Allergy and Infectious Diseases (NIAID), part of the National Institutes of Health (NIH), recently awarded an $11.9 million contract to the University of Rochester (UR) Medical Center to continue the work of the Center for Biodefense Immune Modeling (CBIM) for another five years. The co-directors of the UR-CBIM are Dr. Hulin Wu, a statistician and Professor of Biostatistics (and IMS member), Department of Biostatistics and Computational Biology, and Dr. Martin Zand, a professor in the Department of Medicine at the UR Medical Center. The Center provides a good example for successful collaborations between quantitative scientists and experimental investigators.


This successfully awarded renewed proposal builds on the work of the UR-CBIM, which was established five years ago, with the original total award of $10 million. During the past five years, the UR-CBIM, composed of a multi-disciplinary team of statisticians, mathematical and engineering modelers, computer scientists, software developers, bioinformaticians, and biomedical investigators, sought to develop novel mathematical models, statistical methods, and bioinformatics tools to decipher the host immune responses to influenza infection and other infectious agents. Extensive experimental data from mice were collected to support the development and validation of the proposed models, statistical methods, and bioinformatics tools. This project resulted in 60 manuscripts (50 published/accepted and 10 submitted or under revision) over the past five years: 10 in biomathematical modeling journals, 27 in statistical or biostatistical journals, 5 in bioinformatics/biocomputing journals, and 18 in biomedical science journals. Many of these papers were published in top statistical/biostatistical journals such as the Annals of Statistics, JASA, Biometrika, and Biometrics. Several bioinformatics tools were also developed. In particular, a differential equation modeling tool, DEDiscover, is a user-friendly software for both modelers and biomedical scientists, providing model simulation, parameter estimation and model evaluation functions, freely available at https://cbim. urmc.rochester.edu/software. Extensive experimental data for model development and model validation were also generated and will be available via the NIAID/NIH immunology database, ImmPort (www.immport.org) and the UR-CBIM website.


The UR-CBIM aims to develop multi-level and multi-type mathematical/computational models for immune responses that include ordinary differential equation models, stochastic differential equation models, state-space models, stochastic process models, agent-based models, and network models at the cellular, protein, and genetic levels. Novel statistical methods will be developed for these complex dynamic models that are used to quantify immune responses to influenza infection and vaccination. Extensive experiments in both mice and human subjects will be performed and high-throughput experimental data at cellular, protein, and genetic levels will be collected to support development, validation, and testing of the proposed models and methods. The developed modeling techniques and statistical methods will be converted into user-friendly tools for biomedical investigators and modelers to easily use to further better understanding of immune responses and design of new experiments for discovering new immune-based therapies and vaccines against infectious agents. The collaboration and interaction among multi disciplinary investigators, in particular between quantitative modelers and experimental scientists, has been and will continue to be crucial for the success of this project.


The University of Rochester is one of the four institutes selected to establish Centers for Modeling Immunity for Biodefense by NIH.The University of Rochester’s CBIM project, led by Drs. Wu and Zand, consists of five cores: Mathematical Modeling Core,
led by Drs. Alan Perelson and Hongyu Miao; Statistics Core, led by Dr. Wu; a Bioinformatics Core, led by Drs. Gregory Warnes and Jingming Ma; Immunology Core, led by Dr. Zand; and an Education Program, led by Drs. Hua Liang and Tim Mosmann. Dr. Alan Perelson, a senior fellow at the Los Alamos National Lab and
a world leader in immune modeling, is the subcontract PI of the UR-CBIM project.


Statisticians, like Dr. Wu, play critical roles in successful biomedical research, but rarely do they lead the large multidisciplinary research projects their work supports. This award demonstrates the importance of equally using statistical and mathematical modeling approaches, and not solely experimental methods, to make scientific discoveries. Dr. Wu joined the University of Rochester in 2003 and founded the interdisciplinary Division of Biomedical Modeling and Informatics at the Department of Biostatistics and Computational Biology with a goal of integrating multi-disciplinary quantitative and computational sciences for biomedical research.


Besides this Biodefense Center project, the Division, led by Dr. Wu, also provides statistics, mathematical modeling and bioinformatics support to several large centers or program projects at the University of Rochester Medical Center.