Center for Statistics in Big Data (CSBD)
Advancing statistical and computational methods to transform high-dimensional multi-modal biomedical data into actionable insights for precision health.
CSBD Research
The Center for Statistics in Big Data (CSBD) develops innovative statistical and artificial intelligence methods for analyzing complex biomedical and population health data. Our research combines methodological rigor with interdisciplinary collaboration to advance biological discovery, precision health, and evidence-based decision-making.
Focus Areas
Statistical Foundations of Biomedical AI
We develop statistical methods that make artificial intelligence more reliable, interpretable, robust, and generalizable for biomedical and health applications. Areas of emphasis include uncertainty quantification, calibration, transfer learning, generative models, and valid statistical inference using AI-generated and external data.
Population-Level and Multimodal Biology
We develop methods for studying biological variation across individuals and populations using complex, multimodal data. Our work integrates molecular, cellular, imaging, clinical, environmental, and behavioral information to characterize biological systems at multiple scales.
Genomics, Single-Cell and Spatial Omics
We develop statistical and computational methods for understanding genetic regulation, cellular heterogeneity, tissue organization, and disease mechanisms. Applications include population genomics, single-cell and spatial data analysis, multi-omics integration, genetic association studies, and perturbation experiments.
Microbiome, Virome and Host–Microbe Systems
We develop methods to identify, quantify, and characterize microbial and viral communities and their interactions with human hosts. Our research addresses microbial community structure, functional annotation, community dynamics, host–microbe interactions, and the integration of metagenomic, metabolomic, and clinical data.
Precision Health and Learning Health Systems
We integrate electronic health records, molecular profiles, medical imaging, wearable-device data, and longitudinal health information to improve prediction, disease monitoring, and clinical decision-making. Our goal is to support personalized care while enabling health systems to learn continuously from real-world data.
Causal and Translational Data Science
We develop methods for causal inference, data integration, heterogeneous treatment-effect estimation, and principled study design. Through close collaboration with biomedical and clinical investigators, we seek to translate methodological advances into improved scientific evidence, clinical practice, and population health.
Publications
Selected Publications from CSBD Researchers
Bittinger K, Zhao C, Li Y, Ford E, Friedman ES, Ni J, Kulkarni CV, Cai J, Tian Y, Liu Q, Patterson AD, Sarkar D, Chan SHJ, Maranas C, Saha-Shah A, Lund P, Garcia BA, Mattei LM, Gerber JS, Elovitz MA, Kelly A, DeRusso P, Kim D, Hofstaedter CE, Goulian M, Li H, Bushman FD, Zemel BS, Wu GD. Bacterial colonization reprograms the neonatal gut metabolome. Nature Microbiology, 2020 Jun;5(6):838-847. doi: 10.1038/s41564-020-0694-0.
Bushman FD, Conrad M, Ren Y, Zhao C, Gu C, Petucci C, Kim MS, Abbas A, Downes KJ, Devas N, Mattei LM, Breton J, Kelsen J, Marakos S, Galgano A, Kachelries K, Erlichman J, Hart JL, Moraskie M, Kim D, Zhang H, Hofstaedter CE, Wu GD, Lewis JD, Zackular JP, Li H, Bittinger K, Baldassano R. Multi-omic Analysis of the Interaction Between Clostridioides difficile Infection and Pediatric Inflammatory Bowel Disease. Cell Host Microbe. 2020 Sep 9;28(3):422-433.e7. doi: 10.1016/j.chom.2020.07.020.
Cao Y, Zhang A, Li H. Multisample Estimation of Bacterial Composition Matrices in Metagenomics Data. Biometrika. Oxford University Press (OUP); 2020 Mar 1;107(1):75–92.
Cullen CM, Aneja KK, Beyhan S, Cho CE, Woloszynek S, Convertino M, McCoy SJ, Zhang Y, Anderson MZ, Alvarez-Ponce D, Smirnova E, Karstens L, Dorrestein PC, Li H, Sen Gupta A, Cheung K, Powers JG, Zhao Z, Rosen GL. Emerging Priorities for Microbiome Research. Frontiers in Microbiology. 2020 Feb 19;11:136. doi: 10.3389/fmicb.2020.00136.
Ma R, Cai TT, Li H. Global and Simultaneous Hypothesis Testing for High-Dimensional Logistic Regression Models. J Am Stat Assoc. 2021;116(534):984-998. doi: 10.1080/01621459.2019.1699421.
Ma R, Cai TT, Li H (2020): Optimal permutation recovery in permuted monotone matrix model. Journal of the American Statistical Association, accepted.
Ma R, Cai TT, Li H. Optimal Permutation Recovery in Permuted Monotone Matrix Model. J Am Stat Assoc. 2021;116(535):1358-1372. doi: 10.1080/01621459.2020.1713794.
Ma R, Cai TT, Li H (2020): Optimal estimation of bacterial growth rates based on permuted monotone matrix. Biometrika, accepted.
Sheng Z, Qiu C, Liu H, Gluck C, Hsu J, He J, Hsu CY, Sha D, Weir MR, Isakova T, Raj DS, Ricon-Choles H, Feldman HI, Townsend R, Li H, Susztak K (2020): Integrated genetic-epigenetic analysis supports a causal role for inflammation in diabetic kidney disease pathogenesis. Proceedings of the National Academy of Sciences, accepted.
Wang S, Cai TT, Li H (2020): Optimal estimation of Wasserstein distance on a tree with an application to microbiome studies. Journal of the American Statistical Association, accepted.
Wang S, Cai TT, Li H (2020): Hypothesis testing for phylogenetic composition: A minimum-cost flow perspective. Biometrika, accepted.
Yarmarkovich M, Farrell A, Sison A III, Di Marco M, Raman P, Parris J, Monos DS, Lee H, Stevanovic S, Maris JM (2020): Immunogenicity and immune silence in human cancer. Frontiers in Immunology, 11:69.