HapCluster

HapCluster performs haplotype clustering and Markov-chain Monte Carlo (MCMC)-based fine-mapping to detect causal loci in SNP-based association studies.

Key Features:

• Between-Haplotype Scoring: Implements a between-haplotype score that captures evolutionary relatedness among haplotypes near an unobserved causal locus.
• Haplotype Clustering: Defines clusters of haplotypes based on a putative ancestral haplotype and a specified cut-off distance.
• Genotype Prediction: Predicts the genotype at the causal locus by counting how many of an individual's two haplotypes fall within defined clusters.
• MCMC Implementation: Uses a Markov-chain Monte Carlo algorithm to search across genomic locations and ancestral haplotypes for clusters enriched with case data.
• Empirical Accuracy: Demonstrated nearly 98% accuracy in predicting genotypes at the causal locus in test analyses with real-world data.
• Comparative Performance: Shown in simulation studies to outperform alternative methods, particularly where multi-point approaches improve on single-SNP analyses.
• Computational Efficiency and Scope: Designed to run efficiently across diverse disease models and genomic scales for fine-mapping applications.

Scientific Applications:

• Genetic association studies: Fine-map causal mutations contributing to phenotypic traits using SNP and haplotype data.
• Fine-mapping of causal loci: Localize and prioritize candidate causal variants through cluster enrichment analysis.
• Evaluation of multi-point methods: Compare multi-point haplotype-based analyses with single-SNP approaches in simulation and empirical studies.
• Investigation of genetic architecture: Explore haplotype structure and evolutionary relatedness around putative causal loci to inform disease genetics.

Methodology:

Computes a between-haplotype score, forms haplotype clusters around a putative ancestral haplotype with a cut-off distance, predicts genotypes by counting individual haplotypes within clusters, and implements a Markov-chain Monte Carlo algorithm to search locations and ancestral haplotypes for case-enriched clusters.