Jasmine

Jasmine merges structural variant (SV) calls from long-read sequencing datasets into harmonized multi-sample callsets to enable accurate population- and trio-level SV analysis.

Key Features:

• Automated alignment and SV calling: Performs automated alignment and structural variant calling on long-read sequencing data.
• SV merging across samples: Represents SVs as a network and applies a modified minimum spanning forest algorithm to merge equivalent variants across samples.
• Proximity graph for SV comparison: Constructs an SV proximity graph to compare variants, improving accuracy and efficiency and outperforming five widely-used methods.
• Reduction of Mendelian discordance: Reduces Mendelian discordance in trio datasets by more than five-fold.
• Identification of high-confidence de novo variants: Identifies high-confidence de novo structural variants confirmed using multiple long-read sequencing technologies.
• Harmonized callset generation: Produces a harmonized callset of 205,192 SVs from 31 long-read-sequenced samples of diverse ancestry.
• Genotyping and impact assessment on gene expression: Genotypes SVs in 444 short-read samples from the 1000 Genomes Project using DNA and RNA sequencing data and assesses variant impacts on gene expression, including medically relevant genes.

Scientific Applications:

• Population genetics: Enables comparative analyses of SV distributions across populations using harmonized multi-sample callsets.
• Genomic medicine: Supports identification and validation of medically relevant and de novo SVs for disease studies.
• Evolutionary biology: Facilitates study of population-specific variant distributions and evolutionary patterns of structural variation.
• Gene expression regulation: Enables assessment of SV effects on gene expression through DNA- and RNA-based genotyping.

Methodology:

Performs alignment and SV calling on long-read data, represents SVs as a network and constructs an SV proximity graph, merges variants using a modified minimum spanning forest algorithm, and genotypes SVs in short-read samples using DNA and RNA sequencing data.