Figure 5

Analysis of HIV proteomes. Each protein sequence is plotted on the horizontal axis according to its size (dots). Red dots are significant (p < 0.001). Above each large cloud of sequences, the corresponding protein name is denoted. Sequences in between clouds are likely to be truncated version of larger proteins that were not pruned from the data set. The grey line denotes the average score. (A) The CBP method yielded only 153 out of 11017 sequences whose epitope distribution was unlikely to be random (less than 2%). (B) The H&S index yielded 68 sequences unlikely to be random (less than 1%). Marked with green dots are the predicted ancestral sequence of HIV-1 clade B proteins.