Principal component analysis of the motif representation data of HIV-1 sequences from subtypes/recombinant B, C, A1 and 01_AE.

<p>a) Scores plot (PC3 <i>vs</i>. PC4). Each point is an HIV-1 full genome sequence. Hypermutated sequences identified by the LANL database are indicated by “H” <a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0087679#pone.0087679-Ebrahimi2" target="_blank">[11]</a>. b) Loadings plot (PC3 <i>vs</i>. PC4). Each point is a motif. The motifs associated with hypermutation by APOBEC3G appear as outliers. Note that only two sequences (indicated by arrows) contain GA→AA footprints that is a hallmark of mutation by APOBEC3F and/or other similar APOBEC3 enzymes <a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0087679#pone.0087679-Ebrahimi2" target="_blank">[11]</a>.