Strobealign: flexible seed size enables ultra-fast and accurate read alignment

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Read alignment is a critical and computationally intensive step in next-generation sequencing analysis.
• Existing seeding methods for sequence comparison have limitations in speed and handling repetitive sequences.
• Advances in seeding algorithms like syncmers and strobemers offer potential for faster alignment.

Purpose of the Study:

• To develop a novel seeding approach for fast sequence comparison by combining syncmers and strobemers.
• To implement this novel seeding method in a short-read aligner named strobealign.
• To evaluate the performance of strobealign in terms of speed and accuracy compared to existing aligners.

Main Methods:

• Developed a new seeding strategy using a combination of syncmers and strobemers to create dynamic-sized fuzzy seeds.
• Implemented the novel seeding approach in a short-read alignment tool, strobealign.
• Introduced a new metric, E-hits, to quantify and reduce repetitiveness during the seeding process.

Main Results:

• The novel seeding approach effectively reduces repetitiveness in the seeding step, as measured by E-hits.
• strobealign demonstrates several-fold speed improvements over traditional aligners.
• strobealign achieves similar or higher accuracy compared to traditional aligners and outperforms recent short-read aligners for reads >= 150nt.

Conclusions:

• The combination of syncmers and strobemers provides an effective and fast seeding strategy for short-read alignment.
• strobealign offers a significant advancement in computational efficiency and accuracy for large-scale genomic analyses.
• The developed method addresses the computational bottleneck in read alignment, making genomic analyses faster and more accessible.