Fast sequence alignment for centromeres with RaMA

Area of Science:

• Genomics
• Bioinformatics

Background:

• The human pangenome enables study of complex genomic regions like centromeres, which contain extra-long tandem repeats (ETRs).
• Existing sequence alignment tools struggle to accurately represent genetic variations within ETRs.

Purpose of the Study:

• To develop an advanced alignment tool, rare match aligner (RaMA), for improved analysis of complex genomic regions.
• To enhance the accuracy and efficiency of sequence alignment for centromeric DNA.

Main Methods:

• Developed RaMA using rare matches as anchors and a two-piece affine gap cost model.
• Implemented parallel computing and the wavefront algorithm for accelerated anchor discovery and alignment.
• Introduced novel methods for defining reliable alignment regions.

Main Results:

• RaMA achieved more accurate alignments compared to existing methods, effectively capturing higher-order repeat (HOR) structures in simulated and real data (CHM13, CHM1).
• The tool demonstrated significant speed improvements (up to 13.66x faster) and reduced memory usage (11% of UniAligner).
• Enhanced accuracy in centromeric alignment statistics through refined alignment region definitions.

Conclusions:

• RaMA provides a more accurate and efficient solution for aligning complex genomic sequences, particularly in centromeric regions.
• The tool advances the analysis of genetic events within ETRs, crucial for understanding human pangenome diversity.
• RaMA represents a significant improvement for genomic research involving repetitive DNA elements.