Learning Resistance Mutation Pathways of HIV

We propose a novel machine learning algorithm for learning mutation pathways of viruses from a population of viral DNA strands. More specifically, given a number of sequences, the algorithm constructs a phylogenetic tree that expresses the ancestry of the sequences, and at the same time builds a model describing dependencies between mutations that is consistent with the data as well as the phylogenetic tree. Our approach extends existing approaches for phylogenetic tree construction by not assuming mutations to be independent, and extends existing approaches for detecting dependencies between different positions of a sequence by not assuming the sequences to be independent.