IsoAligner: dynamic mapping of amino acid positions across protein isoforms

Abstract

Aligning protein isoform sequences is often performed in cancer diagnostics to homogenise mutation annotations from different diagnostic assays. However, most alignment tools are fitted for homologous sequences, leading often to alignments of non-identical exonic regions. Here, we present the interactive alignment webservice IsoAligner for exact mapping of exonic protein subsequences. The tool uses a customized Needleman-Wunsch algorithm including an open gap penalty combined with a gene-specific minimal exon length function and dynamically adjustable parameters. As an input, IsoAligner accepts either various gene/transcript/protein IDs from different databases (Ensembl, UniProt, RefSeq) or raw amino acid sequences. The output of IsoAligner consists of pairwise alignments and a table of mapped amino acid positions between the canonical or supplied isoform IDs and all alternative isoforms. IsoAligner’s human isoform library comprises of over 1.3 million IDs mapped on over 120,000 protein sequences. IsoAligner, is a fast and interactive alignment tool for retrieving amino acids positions between different protein isoforms. Its application will allow diagnostic and precision medicine labs to detect inconsistent variant annotations between different assays and databases. Availability: This tool is available as a Webservice on www.isoaligner.org. A REST API is available for programmatic access. The source code for both services can be found at https://github.com/mtp-usz/IsoAligner.