HIVCoR

HIVCoR predicts coreceptor usage of HIV-1 CRF01_AE sequences to inform antiretroviral therapy selection, including consideration of coreceptor-specific inhibitors.

Key Features:

• Algorithmic Approach: Implements random forest and support vector machines (SVM) for classification of coreceptor usage.
• Input Features: Utilizes amino acid compositions, pseudo amino acid compositions, and relative synonymous codon usage frequencies as predictive features.
• Validation and Performance: Achieved a 93.79% overall success rate in external validation on a CRF01_AE benchmark dataset, indicating high accuracy relative to genotypic predictors developed predominantly for subtypes B and C.

Scientific Applications:

• Antiretroviral therapy selection: Supports selection of coreceptor-targeted inhibitors by predicting viral coreceptor usage for CRF01_AE infections.
• Clinical and research decision support: Provides genotype-based information for clinicians and researchers studying tropism and treatment strategies for the CRF01_AE subtype.

Methodology:

Combines random forest and SVM models trained on features including amino acid composition, pseudo amino acid composition, and relative synonymous codon usage frequencies.