GERMLINE

GERMLINE identifies segmental sharing indicative of recent common ancestry between pairs of individuals in large whole-genome cohorts.

Key Features:

• Scalability: Scales linearly with sample size, enabling analysis of large cohorts and whole-genome data.
• Haplotype dictionary: Constructs a dictionary of haplotypes to detect short exact matches between individuals' genomic data.
• Dynamic programming expansion: Uses dynamic programming to extend short exact matches into long, nearly identical segments.
• Concordance with other methods: Demonstrates high concordance with other methods capable of processing similar data.

Scientific Applications:

• Resolution of phasing errors: Aids identification and resolution of phasing errors by analyzing hidden relatedness.
• Detection of polymorphic deletions: Exposes polymorphic deletions that are challenging to detect with other methods, with validation via concordance with independent databases and statistical analyses (e.g., fluorescence intensity), although GERMLINE does not directly use fluorescence intensity.
• Population-scale surveys: Applied to surveys of hidden relatedness in datasets such as HapMap and a densely typed island population of ~3,000 individuals to analyze segmental sharing at scale.

Methodology:

Constructs a haplotype dictionary, identifies short exact matches between individuals, and applies dynamic programming to expand matches into long nearly identical segments, with algorithmic scaling linear in sample size.