Evaluation of Eight Different Bioinformatics Tools to Predict Viral Tropism in Different HIV-1 Subtypes

Hospital Carlos III, Madrid, Spain; Eurofins Viralliance Inc., Kalamazoo, Michigan, USA; Bioalliance Pharma, Paris, France; Hospital Clínico Universitario San Cecilio, Granada, Spain; Hospital de Conxo-CHUS, Santiago de Compostela, Spain

^* To whom correspondence should be addressed. Email: cmendoza{at}teleline.es.

	Abstract

HIV-1 tropism can be assessed using phenotypic assays, but this is quite laborious, expensive, time consuming and only can be made in sophisticated laboratories. More accessible albeit reliable tools for testing HIV-1 tropism are needed in view of the prompt introduction of CCR5 antagonists in clinical practice. Bioinformatics tools based on V3 sequences might help to predict HIV-1 tropism; however, most of these methods have been designed taking into consideration only genetic information derived from HIV-1 subtype B. The aim of this study was to evaluate the performance of several genotypic tools to predict HIV-1 tropism in non-B subtypes, as data available on this issue are scarce. Plasma samples were tested using a new phenotypic tropism assay (PHENOSCRIPT®-tropism, Eurofins), and results were compared with estimates of co-receptor usage using eight different genotypic predictor softwares (SVM, C4.5, C4.5 only with p8-p12, PART, Charge Rule, geno2pheno co-receptor, PSSM_X4R5 and PSSM_sinsi). A total of 150 samples were tested, 115 belonging to patients infected with non-B subtypes and 35 drawn from subtype B-infected patients, which were taken as controls. Testing non-B subtypes, the concordance between the results obtained using the phenotypic assay and distinct genotypic tools was as follows: SVM 78.8%, C4.5 77.5%, C4.5 only with p8-p12 82.5%, PART 82.5%, Charge Rule 82.5%, PSSM_X4R5 82.5%, PSSM_sinsi 83.8%, and geno2pheno 71.3%. Testing clade B viruses, the best concordances were seen for PSSM_X4R5 (91.4%), PSSM_sinsi (88.6%) and geno2pheno (88.6%). The sensitivity for detecting X4 variants was lower testing non-B than B viruses, especially in the case of PSSM_sinsi (38.4% vs 100%), SVM_wetcat (46% vs 100%) and PART (30% vs 90%). In summary, while inference of HIV-1 co-receptor usage using genotypic tools seems to be reliable for clade B viruses, their performance is poor testing non-B subtypes, in which they particularly fail to detect X4 variants.