PROJECT SUMMARY Despite the durable suppression of viral replication by ART, HIV persists indefinitely in infected individuals. Several promising avenues to cure HIV-1 infection have come to light, including gene editing of the CCR5 receptor in CD4 T cells, anti-PD1 monoclonal antibody immunotherapy and the infusion of broadly neutralizing antibodies. These therapeutic approaches constitute a prime opportunity to extensively understand the underlying mechanisms associated with the establishment and maintenance of the HIV reservoir, which will ultimately serve to identify key novel targets for future more refined therapies. Furthermore, the heterogeneity in human immune function has been mapped to multiple environmental factors such as microbiome, metabolome and diet, some of which have been associated with the maintenance of the HIV-1 reservoir. In this P01, we hypothesize that specific key metabolites, microbes and other environmental factors influence the responsiveness to different therapeutic approaches targeting the HIV reservoir. The main objective of the Machine Learning and Modeling Core (MLMC) will be to bring together all datasets generated by Projects 1-3 into a cohesive whole to generate mechanistic models of HIV reservoir maintenance. We shall look into how metabolites modulate the immune transcriptome and epigenome of many subsets. In Aim 1, the MLMC will provide statistical and bioinformatics support for all projects and identify key correlates of HIV viral rebound and HIV DNA decay from all large-scale datasets (OMICs). In Aim 2, the MLMC will perform integrative analysis using novel datasets generated in Aim 1. In Aim 3, the Core will integrate parallel models of regulation of the HIV reservoir into a global unified model where key recurrent features will be identified as prime targets for future therapeutic avenues. The MLMC will thus serve as the central resource for this U19 for the integration of all datasets and generation of mechanistic insights.