RegressHaplo

RegressHaplo: Penalized regression-based haplotype reconstruction for early HIV/SIV infection

RegressHaplo reconstructs viral haplotypes from next-generation sequencing (NGS) data in early HIV and SIV infections by modeling low genetic diversity and convergent evolution across multiple genomic loci using a penalized regression framework.

Key Features:

• Penalized Regression Framework: Applies regression with a penalty term to balance model fit and minimize the number of inferred haplotypes, improving reconstruction accuracy in low-diversity viral populations.
• Early Infection-Specific Algorithm: Accounts for low genetic diversity and convergent evolution patterns characteristic of early HIV/SIV infection in NGS datasets.
• Computational Feasibility in Low Diversity Contexts: Leverages limited sequence variation to enable efficient regression fitting across large viral genomic regions.

Scientific Applications:

• Early HIV/SIV Infection Analysis: Reconstructs viral haplotypes from NGS data collected during initial months of HIV or SIV infection to study viral evolution and adaptation.

Methodology:

Constructs regression models in which candidate haplotypes are treated as covariates, and optimizes an objective function that balances goodness-of-fit to observed sequencing data with a penalty that constrains the total number of inferred haplotypes, enhancing accuracy under low genetic diversity conditions.