A flowchart illustrates procedures for CD4+T cell epitope prediction. (1) Using KDS platform with datasets of 821 circulating HIV-1 strains and 46 HLA-DRB1 alleles in Chinese population; (2) the software predicted possible epitopes by 7 known methods from the IEDB database; (3) all results were combined and scored using and ; (4) the epitopes were ranked according to the score; (5) the epitope with the top score and the lowest IC50 value was selected; (6) the selected epitope was then removed from the epitope pool; (7) steps 4–6 were repeated until all 20 epitopes fulfilled the criteria that were selected.
