IBDmap

IBDmap computes genome-wide identity-by-descent (IBD) sharing from SNP data to distinguish IBD from identity-by-state (IBS) and support identification of causative genetic variants in complex multifactorial disease studies.

Key Features:

• Factorial Hidden Markov Model (HMM): Uses a factorial HMM-based algorithm to compute genome-wide IBD sharing from SNP data.
• Locus-wise pairwise IBD probabilities: Estimates the probability of IBD at each locus for every pair of individuals.
• Enhanced precision and recall: For g-degree relatives (g ≥ 8) achieves over 50% higher precision in IBD tagging compared to previous methods while maintaining ~95% recall.
• Model flexibility: Supports multiple models including a standard model and a 4-track model that incorporates a first-order Markovian model for the linkage disequilibrium (LD) process within subsets of founder populations.
• Efficient computation: Reduces computational complexity by transforming the inheritance-vector state space from exponential to quadratic as a function of g.

Scientific Applications:

• Complex disease variant discovery: Differentiates IBD from IBS to improve identification of genetic variants likely to be causative in complex multifactorial diseases.
• Genetic linkage analysis: Provides accurate locus-wise IBD probability estimates for pairs of individuals to facilitate discovery of disease-associated genomic regions.

Methodology:

Processes SNP data and employs a factorial HMM to estimate locus-wise IBD probabilities, supports a standard and a 4-track model with a first-order Markovian model for LD in founder subsets, and transforms the inheritance-vector state space from exponential to quadratic relative to g.