elife_data3

The files contained in here are the processed data files used to infer the Potts models and annotated as below:

Protease (PR)

• pr.exper.fullseq : Alignment of our processed drug experienced HIV-1 subtype B Protease sequences (obtained from the Stanford HIVDB) in the full 20-letter amino acid.

• pr.exper.reduce4.seq : Alignment of our processed drug-experienced HIV-1 subtype B Protease sequences in the reduced 4-letter (ABCD) amino acid alphabet as used to infer the Potts model.

• pr.exper.weights : Sequence weights of individual sequences in pr.fullseq or pr.reduce4.seq. Sequences are given weights such that the effective number of sequences obtained from a single patient is 1.

Reverse Transcriptase (RT)

• rt.exper.fullseq : Alignment of our processed drug experienced HIV-1 subtype B RT sequences (obtained from the Stanford HIVDB) in the full 20-letter amino acid.

• rt.exper.reduce4.seq : Alignment of our processed drug-experienced HIV-1 subtype B RT sequences in the reduced 4-letter (ABCD) amino acid alphabet as used to infer the Potts model.

• rt.exper.weights : Sequence weights of individual sequences in rt.fullseq or rt.reduce4.seq. Sequences are given weights such that the effective number of sequences obtained from a single patient is 1.

Integrase (IN)

• in.exper.fullseq : Alignment of our processed drug experienced HIV-1 subtype B IN sequences (obtained from the Stanford HIVDB) in the full 20-letter amino acid.

• in.exper.reduce4.seq : Alignment of our processed drug-experienced HIV-1 subtype B IN sequences in the reduced 4-letter (ABCD) amino acid alphabet as used to infer the Potts model.

• in.exper.weights : Sequence weights of individual sequences in in.fullseq or in.reduce4.seq. Sequences are given weights such that the effective number of sequences obtained from a single patient is 1.