BESST--efficient scaffolding of large fragmented assemblies

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• High-throughput sequencing (HTS) is standard for de novo assembly, but short reads hinder contiguity.
• Scaffolding is crucial for improving genome assembly, yet current methods have limitations.
• Existing scaffolding algorithms often rely solely on read pair counts, ignoring contig features and limiting performance on complex genomes.

Purpose of the Study:

• To develop a novel scaffolding algorithm that addresses limitations of existing methods.
• To evaluate the performance of the new algorithm, BESST, against popular stand-alone scaffolders.
• To assess scaffolder performance across a wide range of datasets and genome complexities.

Main Methods:

• Developed BESST, a new algorithm for genome scaffolding.
• Performed comprehensive comparisons of BESST with existing scaffolders on diverse datasets.
• Evaluated scaffolders on datasets of varying sizes and complexities, including the 20 Gbp P. abies genome.

Main Results:

• BESST effectively scaffolds genomes of all sizes and complexities.
• Popular scaffolders showed limitations with complex datasets and no single tool outperformed others universally.
• BESST demonstrated favorable performance on GAGE datasets and outperformed others with wide insert size distributions.

Conclusions:

• Scaffolding can benefit from information sources beyond simple read pair counts.
• BESST offers a robust and versatile solution for genome scaffolding.
• The evaluation highlights the need for comprehensive testing of scaffolding tools on diverse genomic data.