BioMSA

Multiple Sequence Alignment in JavaScript

BioMSA is a JavaScript library for computing alignments of biological sequences (DNA or Protein) locally in the browser. It performs progressive alignments using a weighting scheme similarly to what programs like ClustalW or MUSCLE do. To the best of my knowledge, it is the only multiple sequence alignment library written in JavaScript... and it's also the fastest!

Installing / Getting started

BioMSA is available as a single minified javascript file. Insert the following snippet in your HTML page to start using it.

<script src="https://cdn.jsdelivr.net/npm/biomsa/dist/biomsa.js"></script>

Once loaded, the global biomsa object is available and its align() method can be called to align sequences.

biomsa.align(["ACTGGGGAGGTGTA", "ACTGAGGTGTA"]).then((result) => {
  console.log(result);
});

// Array [ "ACTGGGGAGGTGTA", "ACT---GAGGTGTA" ]

Note: align() returns a promise.

NPM package

BioMSA is also available as an NPM package and can be used as an ECMAScript compatible module too. It is shipped with type declarations too.

npm install biomsa

Library options

The align() method has 2 parameters:

• an array of sequences to align
• an optional configuration object

biomsa.align(
    ['SEQVENCE...', 'SEQWANCE...', 'CEQWANSE...'],
    {
        gapopen: -11,
        gapextend: -2,
        matrix: [[.....], [.....], ....],
        method: 'auto',
        type: 'auto',
        gapchar: '-',
        debug: false
    }).then(result => console.log(result))

• gapopen Gap open penalty (a negative number). If not provided, it is set based on the sequence type.

• gapextend Gap extend penalty (a negative number). If not provided, it is set based on the sequence type.

• matrix Substitution score matrix as an array of number array. Cells are sorted by amino acid 1 letter code rank (A, C, D, E,...). If not provided, it is set based on the sequence type.

• method (default "auto") Alignment method. By default, the method is set based on the sequences length. For sizes greater than 1600 residues, the diagonal based heuristics is used. For shorter sequences a complete Needleman-Wunsch alignment is performed.

  • "auto" default value
  • "complete" computes an optimal alignment using Needleman-Wunsch algorithm. This can be slow and take a lot of memory for long sequences.
  • "diag" computes an alignment by first finding common segments between sequences (called diagonals) and then computing the missing segments using NW algorithm.

• type (default "auto") Sequence type. Can be "amino", "nucleic" or "auto" when auto-detected. BioMSA encodes non-canonical residues randomly. For example 'B' in a protein sequence which could be "Asn" or "Asp" will be encoded randomly as one of these amino-acids.

• gapchar (default "-") Character to use in the aligned sequences to represent a gap.

• debug (default false) Boolean. Set to true to report some debugging information to the javascript console.

Features

• Multiple sequence alignment of nucleic and proteic sequences.
• Tree guided progressive alignment using a weighing scheme, substitution matrices, alignment score optimization by dynamic programming.
• Approximative fast alignment of large DNA sequences (e.g. x5 16kbases mitochondrial DNA sequences in 100ms), using minimizers and diagonals extension.

Licensing

"The code in this project is licensed under MIT license."