Essential Regression: A generalizable framework for inferring causal latent factors from multi-omic datasets

• Type of Material: Article
• Authors: Xin Bing, Cornell University; Tyler Lovelace, University of Pittsburgh; Florentina Bunea, Cornell University; Marten Wegkamp, Cornell University; Sudhir Kasturi, Emory University; Harinder Singh, University of Pittsburgh; Panayiotis V Benos, University of Pittsburgh; Jishnu Das, University of Pittsburgh
• Language: English
• Date: 2022-05-13
• Publisher: RELX
• Publication Version: Final Published Version
• Copyright Statement: © 2022 The Author(s)
• License: Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International
• Final Published Version (URL): http://dx.doi.org/10.1016/j.patter.2022.100473
• Title of Journal or Parent Work: Patterns
• Volume: 3
• Issue: 5
• Start Page: 100473
• End Page: 100473
• Grant/Funding Information: This study was partially supported by NIH grants DP2AI164325 to J.D., U01HL137159, R01HL140963, R01HL159805, and R01HL157879 to P.V.B., U01AI141990 to H.S., and F31LM013966 to T.L.; NSF grants DMS-1712709 and DMS-2015195 to F.B. and M.W.; and DoD grant W81XWH2110864 to J.D. H.S. also acknowledges support from the UPMC ITTC fund. S.P.K. acknowledges support from the Yerkes Pilot Research Pilot Program (part of the Yerkes NPRC Base Grant, P51-OD011132).
• Keywords: interpretable machine learning; latent model; systems biology; causal inference; dimensionality reduction; machine learning; systems immunology
• Research Categories: Health Sciences, Immunology; Biology, Biostatistics

Abstract

High-dimensional cellular and molecular profiling of biological samples highlights the need for analytical approaches that can integrate multi-omic datasets to generate prioritized causal inferences. Current methods are limited by high dimensionality of the combined datasets, the differences in their data distributions, and their integration to infer causal relationships. Here, we present Essential Regression (ER), a novel latent-factor-regression-based interpretable machine-learning approach that addresses these problems by identifying latent factors and their likely cause-effect relationships with system-wide outcomes/properties of interest. ER can integrate many multi-omic datasets without structural or distributional assumptions regarding the data. It outperforms a range of state-of-the-art methods in terms of prediction. ER can be coupled with probabilistic graphical modeling, thereby strengthening the causal inferences. The utility of ER is demonstrated using multi-omic system immunology datasets to generate and validate novel cellular and molecular inferences in a wide range of contexts including immunosenescence and immune dysregulation.