ABSTRACT: CORE B -- COMPUTATIONAL MODELING & ANALYSIS (Core Leader: Lauffenburger) A key premise underlying the motivation for this program is that there exist multiple molecular and cellular features, in both host and pathogen, that operate integratively via complex cross-talks and feedbacks to govern TB disease outcomes. This premise is distinct from univariate perspectives that seek individual features predictive of disease state, and from traditional multivariate perspectives that while admitting effects of numerous features assume that they operate independently. An associated premise is that heterogeneity is crucial to comprehend in both animal and human studies. These considerations compel computational analysis and modeling approaches prioritizing incorporation of multiple features, represented by disparate data types, acting in coordinated manner and exhibiting different quantitative contributions under disparate circumstances and in diverse subpopulations. The Computational Modeling & Analysis Core (CMAC) will apply a spectrum of state-of- art methods, including a number of machine learning techniques, to help address the questions posed in the Projects, in close partnership with the Project investigators. Activity 1. Curation, quality control, and feature identification from RNAseq measurements · Bulk RNAseq from mammalian tissue and blood · Single-cell RNAseq from mammalian lung and blood cells · Barcode and genotype identification for I-Mac library from mouse strains · RNAseq from Mtb strains Activity 2. Modeling relationships between host immune response properties and Mtb infection outcomes · Correlation methods · Unsupervised modeling methods · Supervised modeling methods · Cross-species translation modeling methods Activity 3. Modeling relationships between Mtb genotype, state, and infection outcomes · Correlation methods · Unsupervised modeling methods · Supervised modeling methods · Network modeling methods