Swalo

Swalo performs likelihood-based scaffolding of genome assemblies by ordering and orienting contigs using likelihood models derived from second-generation sequencing reads.

Key Features:

• Likelihood-Based Scaffolding: Swalo employs a generative model for sequencing data to calculate maximum likelihood estimates, determine optimal gaps between contigs, and assess whether linking specific contigs increases the overall assembly likelihood.
• Efficient Implementation: The method incorporates approximations to improve computational efficiency, enabling application to large genomic datasets.
• Performance and Accuracy: Comparative analyses on real and simulated datasets report high accuracy in scaffold construction, producing correct joins at rates comparable or superior to other scaffolding approaches while minimizing incorrect linkages.

Scientific Applications:

• Genome assembly improvement: Enhances scaffold construction to increase the contiguity and overall quality of genome assemblies.
• Gene annotation: Reduces assembly fragmentation to support more accurate gene model prediction and annotation.
• Comparative genomics and evolutionary studies: Provides higher-quality assemblies for comparative analyses and evolutionary inference.

Methodology:

Swalo uses a generative statistical model of sequencing data to compute likelihoods and maximum likelihood estimates for gap sizes and contig links; contigs are linked when joins are unambiguous or when the increase in assembly likelihood from their linkage is significantly greater than alternative connections; approximations are applied to improve computational efficiency.