Blossoc

Blossoc performs block-association mapping by constructing local phylogenetic trees to localize disease-causing genetic variants in high-density case-control association studies and related genome-wide analyses.

Key Features:

• Phylogenetic Tree-Based Analysis: Constructs "perfect" phylogenetic trees around each marker, treating them as decision trees to classify disease status and exploit informative regions near causal mutations.
• Handling of Haplotype and Genotype Data: Analyzes haplotype data and phased genotype data and permits inclusion of incompatible markers to extend regions for phylogeny estimation when markers are sparse.
• Efficiency and Speed: Enables genome-wide analyses of up to 3 million SNPs across 1,000 cases and controls in under two CPU hours.
• Accuracy in Complex Scenarios: Maintains accuracy comparable to single marker association in simple mutation scenarios and outperforms fast data-mining approaches such as HapMiner and Haplotype Pattern Mining (HPM) in settings with mutation heterogeneity and complex haplotype structure (e.g., HapMap data).
• Empirical Validation: Has localized known susceptibility variants such as DeltaF508 (cystic fibrosis) and identified significant associations for traits linked to CYP2D6.
• QBlossoc Extension: QBlossoc extends Blossoc for linkage disequilibrium mapping of quantitative traits by constructing local genealogies and identifying significant clustering of trait values, improving localization and ranking of true positives relative to single-marker methods at contemporary marker densities.

Scientific Applications:

• Genome-wide association studies (GWAS): Localizing disease-causing variants in high-density case-control GWAS.
• Quantitative trait mapping: Linkage disequilibrium mapping and localization of quantitative trait loci using QBlossoc.
• Fine-mapping in complex regions: Fine-mapping and localization in regions with mutation heterogeneity or intricate haplotype patterns (e.g., HapMap-derived structures).

Methodology:

Builds local genealogies by constructing "perfect" phylogenetic trees around markers, uses linkage disequilibrium information between markers, allows inclusion of incompatible markers when marker density is low, and detects significant clustering of case chromosomes or quantitative trait values within these trees.