A rank-based sequence aligner with applications in phylogenetic analysis.

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Short DNA read alignment tools balance correctness and speed.
• Existing methods may not achieve optimal accuracy, especially at low coverage.
• Accurate read assignment is crucial for genomic analysis and organism identification.

Purpose of the Study:

• Introduce a novel rank-based DNA read alignment method using Local Rank Distance (LRD).
• Improve alignment correctness over speed, while investigating indexing strategies for efficiency.
• Evaluate the LRD aligner's accuracy and performance against state-of-the-art tools.

Main Methods:

• Developed a rank-based aligner implementing the Local Rank Distance (LRD) metric.
• Optimized speed by storing k-mer positions in hash tables for each read.
• Introduced an approximate LRD aligner by considering only likely positional matches in the reference genome.

Main Results:

• The LRD aligner achieves high biological accuracy, outperforming other tools.
• Demonstrated superior accuracy even at very low base coverage.
• Experiments confirmed high precision and recall, even with contaminated reads.

Conclusions:

• The LRD aligner is a highly accurate alternative to standard alignment tools, particularly when accuracy is paramount.
• Its effectiveness at low coverage makes it suitable for diverse genomic applications, including organism identification.
• Source code is available for further development and use.