LARDER BA, KEMP S, HERTOGS K; Interscience Conference on Antimicrobial Agents and Chemotherapy.

Abstr Intersci Conf Antimicrob Agents Chemother Intersci Conf Antimicrob Agents Chemother. 2000 Sep 17-20; 40: 306.

Virco UK, Cambridge, United Kingdom

BACKGROUND: Genotypic assays are more rapid and less complex than phenotyping to perform but the information generated is difficult to interpret for clinical practice. Objectives: To develop a rapid, automated, system to generate quantitative predictions of HIV-1 phenotypic drug resistance from a genetic sequence. To assess the accuracy of the virtual phenotype (vPT) predictions and compare them to a rules-based interpretation system.METHODS: A proprietary software system was developed to search our relational database for samples with matching patterns of mutations to the sample being assessed, retrieve the phenotypes from these matching samples and calculate the mean change in IC[50], and the range, for those matched phenotypes: the VirtualPhenotype[TM]. Linear regression analysis was used to compare vPTs with actual phenotypes, for 5000 randomly selected samples. The same genotypic samples were also interpreted according to the DAP (resistance collaborative group) table and compared with actual phenotypes.RESULTS: Comparisons of actual with virtual fold resistance by linear regression gave an overall r[2] value of 0.72 for all samples (>70,000 individual phenotypic results). The PIs gave r[2] values ranging from 0.52 (APV) to 0.78 (RTV). AZT, 3TC and EFV gave r[2] values ranging from 0.64 to 0.84. When the phenotypes were divided into two susceptibility categories (S=<4-FOLD, R=>10-fold), the DAP analysis could under call sensitivity (e.g. for ddI sensitive samples, DAP=50% resistant, vPT=1%). Furthermore, a significant proportion of nelfinavir resistant isolates were called sensitive by the DAP analysis (22% versus 4% for vPT). The vPT was also better than the DAP at identifying d4T resistance.CONCLUSIONS: These results demonstrate a high degree of concordance between virtual and actual phenotypes and indicate that rules-based genotype interpretation systems such as the DAP table can give inaccurate information.KEYWORDS: Phenotypes; Resistance; Virtual

Publication Types:

• Meeting Abstracts

Keywords:

• Case-Control Studies
• Didanosine
• Drug Resistance
• Drug Resistance, Multiple, Viral
• Drug Resistance, Viral
• Genotype
• HIV-1
• Lamivudine
• Mutation
• Nelfinavir
• Phenotype
• Zidovudine
• genetics

Other ID:

• GWAIDS0009652

UI: 102247150

From Meeting Abstracts