Machine learning identified genetic features associated with HIV sequences in the monocytes

Abstract

Multiple alignments of codon and protein sequences were produced using Gene Cutter (https://www.hiv.lanl.gov/content/sequence/GENE_CUTTER/cutter.html). The approximate maximum likelihood phylogenetic trees for env segments were estimated in Molecular Evolutionary Genetics Analysis (MEGA X) using the general time reversible (GTR) + Γ + I nucleotide substitution model (www.megasoftware.net). Support vector machine (SVM), random forest (RF), gradient boosting machine (GBM), extreme gradient boosting with linear booster (XGBL), and extreme gradient boosting with tree booster (XGBT) were applied to compare their accuracy in predicting viruses for monocyte/macrophage. Using previously developed tree-based (SM test) and distance-based (Snn test) tests of compartmentalization, we found evidence for compartmentalization of viral populations from eight individuals between T cells and monocytes [Supplementary Table 2, http://links.lww.com/CM9/B813].