HIV coreceptor tropism determination and mutational pattern identification.

Area of Science:

• Virology
• Computational Biology
• Genetics

Background:

• Human Immunodeficiency Virus Type 1 (HIV-1) initially uses CCR5 for host cell entry.
• Viral tropism can shift to CXCR4 as infection progresses.
• Accurate tropism determination is crucial for effective HIV-1 therapy and understanding viral evolution.

Purpose of the Study:

• To develop a computational classifier for predicting HIV-1 coreceptor usage (CCR5 vs. CXCR4).
• To identify genetic mutation patterns associated with HIV-1 coreceptor switching.
• To provide a tool for monitoring tropism changes in HIV-1 infection.

Main Methods:

• Development of a classifier using two coreceptor-specific weight matrices (CMs) trained on a comprehensive dataset.
• Performance evaluation using ten-fold cross-validation and an independent dataset.
• Analysis of genetic mutation patterns correlated with coreceptor tropism scores.

Main Results:

• The developed classifier achieved high performance: AUC of 0.97, 95.21% accuracy, and 0.885 MCC.
• The classifier outperformed existing methods like geno2pheno and PSSM on an independent dataset.
• Six single-amino acid and three two-amino acid mutational patterns were identified as associated with CXCR4 tropism.

Conclusions:

• The CM-based classifier is a highly accurate tool for predicting HIV-1 coreceptor tropism.
• Identified mutational patterns provide insights into the mechanisms of HIV-1 coreceptor switching.
• The findings aid in selecting appropriate therapies and monitoring disease progression.