LAST

LAST aligns high-throughput DNA sequencing reads by incorporating per-base error probabilities reported by sequencers into its alignment algorithm to improve mapping accuracy.

Key Features:

• Incorporation of per-base error probabilities: LAST uses per-base error probabilities reported by sequencers in its alignment scoring to account for variable sequencing quality.
• Modeling of sequencing errors and genuine variation: The method explicitly models both inherent sequencing errors and genuine sequence differences to distinguish technical errors from biological variation.
• Enhanced mapping accuracy: Integrating error probabilities increases correct mapping rates, exemplified by an increase from 49% to 66% when aligning Drosophila melanogaster reads to the Drosophila simulans genome.
• Applicability to challenging genomic contexts: The approach is suited to data from organisms lacking reference genomes, extinct taxa, or highly polymorphic populations where sequencing errors complicate alignment.

Scientific Applications:

• Comparative genomics of closely related species: Improves alignment accuracy for reads used in cross-species comparisons such as between Drosophila species.
• Analysis of organisms without reference genomes: Enables more reliable use of sequencing reads from taxa that lack assembled reference genomes.
• Studies of extinct or highly polymorphic populations: Facilitates interpretation of reads from degraded, error-prone, or highly variable samples to inform evolutionary and biodiversity research.

Methodology:

Incorporates per-base error probabilities reported by sequencers into the alignment algorithm and explicitly models sequencing errors alongside genuine sequence variation.