[Paper Review] Minimap2: versatile pairwise alignment for nucleotide sequences

Heng Li|arXiv (Cornell University)|Aug 4, 2017

Genomics and Phylogenetic Studies23 citations

TL;DR

Minimap2 is a fast, versatile alignment tool designed for long nucleotide sequences, including ultra-long genomic reads, full-length mRNA, and large contigs. It uses split-read alignment, concave gap penalties, and novel heuristics to achieve 3–4× speedup over short-read mappers and >30× speedup at high accuracy, outperforming specialized aligners across diverse data types.

ABSTRACT

Motivation: Recent advances in sequencing technologies promise ultra-long reads of $\sim$100 kilo bases (kb) in average, full-length mRNA or cDNA reads in high throughput and genomic contigs over 100 mega bases (Mb) in length. Existing alignment programs are unable or inefficient to process such data at scale, which presses for the development of new alignment algorithms. Results: Minimap2 is a general-purpose alignment program to map DNA or long mRNA sequences against a large reference database. It works with accurate short reads of $\ge$100bp in length, $\ge$1kb genomic reads at error rate $\sim$15%, full-length noisy Direct RNA or cDNA reads, and assembly contigs or closely related full chromosomes of hundreds of megabases in length. Minimap2 does split-read alignment, employs concave gap cost for long insertions and deletions (INDELs) and introduces new heuristics to reduce spurious alignments. It is 3-4 times faster than mainstream short-read mappers at comparable accuracy and $\ge$30 times faster at higher accuracy for both genomic and mRNA reads, surpassing most aligners specialized in one type of alignment. Availability and implementation: this https URL Contact: [e-mail]

Motivation & Objective

• To address the growing challenge of aligning ultra-long sequencing reads (up to 100 kb) and large genomic contigs (>100 Mb) that current tools cannot handle efficiently.
• To develop a single alignment tool that performs well across diverse data types, including short reads, long genomic reads, full-length mRNA, and large contigs.
• To improve alignment accuracy for long insertions and deletions (INDELs) using concave gap cost models.
• To reduce spurious alignments through new heuristic filtering strategies.
• To achieve significant speed improvements over existing aligners, especially at high accuracy settings.

Proposed method

• Employs split-read alignment to handle long reads with complex structural variations.
• Applies concave gap cost functions to better model long INDELs, improving alignment accuracy for structural variants.
• Introduces novel heuristics to filter out spurious alignments, reducing false positives.
• Optimizes seed-and-extend heuristics for efficient indexing and alignment of long sequences.
• Supports multiple input types: short reads (≥100 bp), long genomic reads (≥1 kb, ~15% error rate), full-length cDNA/mRNA, and large contigs or chromosomes.
• Uses a generalized algorithm that dynamically adapts to sequence length and error profiles.

Experimental results

Research questions

• RQ1 Can a single alignment algorithm efficiently and accurately map ultra-long reads (up to 100 kb) and large genomic contigs (>100 Mb)?
• RQ2 How can concave gap cost models improve alignment accuracy for long INDELs in noisy long reads?
• RQ3 What heuristics can effectively reduce spurious alignments without sacrificing sensitivity?
• RQ4 To what extent can a unified algorithm outperform specialized aligners in speed and accuracy across diverse data types?
• RQ5 Can the tool maintain high accuracy while achieving orders-of-magnitude speed improvements over existing solutions?

Key findings

• Minimap2 achieves 3–4 times faster performance than mainstream short-read mappers at comparable accuracy levels.
• It is more than 30 times faster than existing tools at higher accuracy settings for both genomic and mRNA reads.
• The tool maintains high alignment accuracy across diverse data types, including noisy full-length cDNA and long genomic reads.
• The use of concave gap costs significantly improves alignment quality for long insertions and deletions.
• The novel heuristics reduce spurious alignments, enhancing specificity without compromising sensitivity.
• Minimap2 outperforms most aligners specialized for one type of alignment, demonstrating broad versatility.