Featherweight long read alignment using partitioned reference indexes.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Nanopore sequencing enables portable genomic research.
• Current long-read aligners and large reference genomes have high memory demands, limiting mobile applications.
• Existing methods for reducing memory usage have limitations.

Purpose of the Study:

• To address the memory constraints of long-read alignment on mobile devices.
• To adapt state-of-the-art aligners for resource-limited computing environments.
• To enable accurate genomic analysis using portable sequencing technologies.

Main Methods:

• Parameter optimization and reference genome partitioning were explored.
• A novel multi-index merging technique was developed and integrated into the Minimap2 aligner.
• Performance was evaluated on a system with 2GB RAM using the human genome.

Main Results:

• Parameter optimization and partitioning showed limitations in reducing memory footprints.
• The integrated multi-index merging technique effectively reduced memory requirements.
• Long-read alignment to the human genome was achieved with negligible accuracy loss on a low-memory system.

Conclusions:

• Multi-index merging is a viable strategy to overcome memory limitations in long-read alignment.
• This approach facilitates portable genomic research by enabling analysis on mobile devices.
• The adapted Minimap2 aligner enhances the accessibility and applicability of Nanopore sequencing data.