@bioscript/seq-align

Sequence alignment algorithms for bioinformatics.

Features

• Global Alignment - Needleman-Wunsch for end-to-end alignment
• Local Alignment - Smith-Waterman for conserved regions
• Semi-Global - End-gap-free alignment for primers/probes
• Overlap Alignment - Suffix-prefix matching for assembly
• Banded Alignment - Fast alignment for >90% identity
• Hirschberg - Memory-efficient O(min(m,n)) space
• 13 Scoring Matrices - BLOSUM, PAM, DNA matrices
• Type Safe - Full TypeScript support
• Zero Dependencies - Pure TypeScript

Installation

npm install @bioscript/seq-align

Quick Start

import { needlemanWunsch, smithWaterman } from '@bioscript/seq-align';

// Global alignment
const global = needlemanWunsch('HEAGAWGHEE', 'PAWHEAE', {
  matrix: 'BLOSUM62',
  gapOpen: -10,
  gapExtend: -1,
});

console.log(global.alignedSeq1);     // 'HEAGAWGHEE'
console.log(global.alignedSeq2);     // '--PAW-HEAE'
console.log(global.identityPercent); // 42.86%

// Local alignment
const local = smithWaterman('HEAGAWGHEEHEAGAWGHEE', 'PAWHEAE', {
  matrix: 'BLOSUM62',
});

console.log(local.alignedSeq1);  // Best matching region
console.log(local.startPos1);    // Start position

API

needlemanWunsch(seq1, seq2, options?)

Global alignment for complete sequences.

Options:

• matrix - Scoring matrix (default: 'BLOSUM62')
• gapOpen - Gap opening penalty (default: -10)
• gapExtend - Gap extension penalty (default: -1)

const result = needlemanWunsch('ACGTACGT', 'ACGTAGCT', {
  matrix: 'DNA_SIMPLE',
  gapOpen: -5,
  gapExtend: -2,
});

smithWaterman(seq1, seq2, options?)

Local alignment for finding best matching regions.

Options: Same as needlemanWunsch, plus:

• minScore - Minimum score threshold (default: 0)

const result = smithWaterman('HEAGAWGHEE', 'AWGHE', {
  matrix: 'BLOSUM62',
  minScore: 20,
});

semiGlobal(seq1, seq2, options?)

Semi-global alignment with free end gaps.

const result = semiGlobal('ATCGATCG', 'GGGGGATCGATCGAAAA', {
  matrix: 'DNA_SIMPLE',
});

overlapAlign(seq1, seq2, options?)

Overlap alignment for sequence assembly.

const result = overlapAlign('ACGTACGTACGT', 'ACGTACGTGGGG', {
  matrix: 'DNA_SIMPLE',
});

bandedAlign(seq1, seq2, options?)

Fast banded alignment for closely related sequences.

Options: Same as needlemanWunsch, plus:

• bandwidth - Half-width of diagonal band (default: 10)

const result = bandedAlign(seq1, seq2, {
  matrix: 'DNA_SIMPLE',
  bandwidth: 10,
});

hirschberg(seq1, seq2, options?)

Memory-efficient global alignment for very long sequences. Uses a linear gap model (gapOpen per gapped column); gapExtend is ignored. Scores are not directly comparable to Needleman-Wunsch when gapOpen !== gapExtend.

const result = hirschberg(longSeq1, longSeq2, {
  matrix: 'DNA_SIMPLE',
});

Scoring Matrices

BLOSUM: BLOSUM45, BLOSUM50, BLOSUM62, BLOSUM80, BLOSUM90

PAM: PAM30, PAM70, PAM120, PAM250

DNA: DNA_SIMPLE, DNA_FULL

import { getMatrix, BLOSUM62 } from '@bioscript/seq-align';

const matrix = getMatrix('BLOSUM62');
const score = BLOSUM62.A.R; // -1

Return Type

All functions return AlignmentResult:

interface AlignmentResult {
  alignedSeq1: string;
  alignedSeq2: string;
  score: number;
  startPos1: number;
  startPos2: number;
  endPos1: number;
  endPos2: number;
  identity: number;
  identityPercent: number;
  alignmentLength: number;
}

License

MIT © 2026 Mykyta Forofontov