HIV Proviral Transcription is Influenced by Multiple Genomic Features Unique to the Integration Site

Abstract

A large body of work has revealed fundamental principles of HIV integration into the human genome, but the effect of the integration site to viral transcription has so far remained elusive. Here, we generate and combine large-scale datasets including epigenetics, transcriptome, and 3D genome architecture to interrogate the chromatin states, transcription activity, and nuclear sub-compartments around HIV integrations in CD4+ T cells to decipher human genome codes shaping HIV expression. Using Hidden Markov and Machine-learning Logistic Regression Models, we defined nuclear sub-compartments and chromatin states contributing to genomic architecture, transcriptional activity, and nucleosome density of regions neighboring the integration site, as additive features influencing HIV expression. Surprisingly, no single feature nor combination of features accurately predicted HIV expression likely reflecting the underlying heterogeneity of HIV placement. While human genes are “precisely” positioned for accurate transcriptional responses, HIV “random” positions dictate the heterogeneous responses as a consequence of the diversity of regulatory features surrounding integration sites.