Figure 3

Data analysis and clustering of V3 loop variant sequences. (a) Histograms for the sets of R5 and DM V3 loop variant sequences based on the number of residue substitutions relative to the native sequence of PDB structure 1ce4. (b) Unsupervised EM clustering of the variant V3 loop sequence dataset based on their residual profile representations yields a cluster purity measure of 90.4%.