Comparison between algorithms analysis of resistance in polymerase region in indels localization of HIV-1 in patients on HAART, attended in health services in São Paulo, Brazil
Presented by Luciana Oliveira Souza (Brazil).
L.O. Souza, G.I.S. Lopez Lopes, K.P. Barroso, J.P.G. Batista, R.B. Andrade, L.F.M. Brígido, Sao Paulo HIV Salvage Workgroup
Instituto Adolfo Lutz, Serviço de Virologia - Laboratório de Genotipagem do HIV, São Paulo, Brazil
Backgroundd: ARV treatment is impaired by viral mutations, among these, insertions and deletions (indels), ranging from 0.5% to 2.5%. The correct location of these mutations is necessary for proper interpretation of the test, as some changes in amino acids (aa) at specific sites may have a major impact on some drugs. The aim of this study is compare algorithms for the identification and localization of these mutations.
Methods: We analyzed 1703 plasma of HIV-1 patients failing therapy, viral load> 2000 copies/mL, assisted by the public health on Sao Paulo State - Brazil, between 2004 and 2009, adopted the standard methodology of Health Ministry. Sequences were edited manually and automated. Subsequently, the sequences were analyzed by the algorithm of Stanford University (http://www.hivdb.stanford.edu/hiv/ - HIV Drug Resistance Database) and Geno2Pheno (http://www.geno2pheno.org/ - Max-Planck-Institute Informatik). For confirmation, the sequences were aligned using the BioEdit program with HXB2.
Results: The use of automatic or manual editing did not show disagreement with each other. We identified 33 indels events. Comparing the algorithms, most of the samples showed position disagreement and sometimes on the aa that were inserted or deleted at both RT and PR. On PR, 7/9 insertions showed disagreement in location (codons 32 to 37). In RT, 15/16 inserts were discordant (identified between codons 64 to 69). For deletions, 6/8 samples were discordant. On BioEdit alignment, all deletions were positioned at 67RT, and not 69RT as indicated by one of the algorithms.
Conclusions: Frequency of indels is similar to described in literature. We observed a significant discrepancy between available algorithms, Stanford and Geno2Pheno. Disagreement about the location can cause deviations on interpretations. The deletion 67RT decreases the susceptibility to NRTIs; different from 69RT, which has unknown impact. Use of current bioinformatics tools should be take into account this limitations, as they may impact on resistance profile.
Back to the session -
Back to the Programme-at-a-Glance