Nebula-kmer

Nebula-kmer genotypes structural variations (SVs) in genomic sequencing data using a mapping-free k-mer counting approach to enable efficient, alignment-free SV genotype inference.

Key Features:

• Mapping-free approach: Uses k-mer counting to eliminate read alignment, providing an order-of-magnitude speed improvement over mapping-based approaches while maintaining comparable accuracy.
• Two-stage workflow: Implements k-mer extraction from known SV coordinates (BED/VCF) using a reference assembly (FASTA) followed by genotyping that counts k-mers in WGS reads and predicts genotypes with a likelihood model.
• SV type coverage: Supports genotyping of indels, deletions, insertions, duplications, inversions, and translocations.
• Generic framework: Represents and genotypes arbitrary structural variation events within a unified framework.
• Reference genotype support: Leverages samples with pre-known genotypes to extract informative k-mers for target SVs.

Scientific Applications:

• Variant cataloging with second- and third-generation sequencing: Facilitates building comprehensive catalogs of common SVs from data generated by second- and third-generation sequencing technologies.
• Large cohort studies: Enables scalable genotyping across large cohorts using whole-genome sequencing (WGS) data.
• Population genetics: Supports population-scale analyses of SV allele frequencies and distribution.
• Disease association studies: Provides genotype calls for association analyses between SVs and phenotypes.
• Evolutionary biology: Allows comparative analyses of structural variation across samples and populations for evolutionary inference.

Methodology:

Extracts k-mers representing input SVs from known coordinates (BED/VCF) using a reference assembly (FASTA) and, where available, samples with pre-known genotypes; counts k-mers in WGS reads and predicts genotypes using a likelihood model by focusing on changes in k-mer counts between reference and sample genomes, thereby bypassing read mapping.