Speaker
Description
Accurately reconstructing the transmitted/founder virus of an HIV infection is challenging. It is particularly difficult to do with heterochronous sequences such as proviruses persisting on ART, because of the challenge of determining an appropriate root position that represents the ancestor. However, since proviral sequences are archived throughout untreated HIV infection and persist long-term during ART without mutating or evolving, we may be able leverage their archival evolutionary history to determine the ancestral sequence. Here, we attempted to estimate an individual's transmitted/founder HIV sequence solely using proviral sequences sampled during ART. We used a data set of 11 participants of a Zambian cohort for whom single-genome-amplified on-ART proviral sequences, as well as plasma HIV RNA sequences collected shortly after seroconversion, were available (median 23; interquartile range 10-25.5 unique proviruses/participant). We inferred phylogenies solely from an alignment of each individual’s proviral sequences. To estimate the position on the tree where the ancestral sequence would be located, we performed outgroup rooting using one of three “closest” Zambian HIV subtype C sequences available on the Los Alamos National Laboratory (LANL) HIV sequence database. With these rooted phylogenies, we evaluated methods to estimate ancestral sequences within phangorn, IQTREE and MrBayes by comparing the inferred ancestral sequences to the consensus of the plasma HIV RNA sequences collected post-seroconversion. We found that using the genetically-closest sequence from LANL to the alignment as the outgroup performed marginally better than using the phylogenetically-closest sequence (either topologically or by substitutions). Overall, we found the different methods comparable with at most 30% error in recovering bases at sites with variation in the proviral alignment in individuals with more than 20 proviral sequences. In summary, we were successful in estimating ancestral transmitted founder viruses solely from proviral sequences sampled on ART. This technique could reveal novel insights into the HIV reservoir dynamics.
