COMPARATIVE PERFORMANCE OF HIGH-DENSITY OLIGONUCLEOTIDE SEQUENCING AND DIDEOXYNUCLEOTIDE SEQUENCING OF HIV TYPE-1 POL FROM CLINICAL-SAMPLES

The performance of the high-density oligonucleotide array methodology (GeneChip) in detecting drug resistance mutations in HIV-1 pol was com pared with that of automated dideoxynucleotide sequencing (ABI) of cli nical samples, viral stocks, and plasmid-derived NL4-3 clones. Sequenc es from 29 clinical samples (plasma RNA, n = 17; lymph node RNA, n = 5 ; lymph node DNA, n = 7) from 12 patients, from 6 viral stock RNA samp les, and from 13 NL4-3 clones were generated by both methods. Editing was done independently by a different investigator for each method bef ore comparing the sequences. In addition, NL4-3 wild type (WT) and mut ants were mixed in varying concentrations and sequenced by both method s. Overall, a concordance of 99.1% was found for a total of 30,865 bas es compared. The comparison of clinical samples (plasma RNA and lymph node RNA and DNA) showed a slightly lower match of base calls, 98.8% f or 19,831 nucleotides compared (protease region, 99.5%, n = 8272; RT r egion, 98.3%, n = 11,316), than for viral stocks and NL4-3 clones (pro tease region, 99.8 %; RT region, 99.5 %). Artificial mixing experiment s showed a bias toward calling wild-type bases by GeneChip. Discordant base calls are most likely due to differential detection of mixtures. The concordance between GeneChip and ABI was high and appeared depend ent on the nature of the templates (directly amplified versus cloned) and the complexity of mixes.