GOparser/index.ts

import fs from 'fs'
import readline from 'readline'

const line$ = (path: string) => readline.createInterface({
    input: fs.createReadStream(path),
    crlfDelay: Infinity
})

// http://geneontology.org/docs/guide-go-evidence-codes/
// http://wiki.geneontology.org/index.php/Category:Evidence_Codes
// http://current.geneontology.org/ontology/go-basic.obo
// http://current.geneontology.org/annotations/goa_human.gaf.gz

const getSymbol = async (symbol:string, goaPath:string, oboPath:string) => {
    const header = [
        'database', 'ID', 'Symbol', 'Qualifier', 
        'GO_Term', 'Evidence', 'Evidence_Code', 
        'With', 'From','Name', 'Alternative_symbols', 
        'Class', 'Taxon', 'Date', 'Origin'
    ]
    const tester = new RegExp('\t'+symbol+'\t')
    const separator = new RegExp('\\|')
    const results = [] as Array<{[key:string]:any}>
    for await (const line of line$(goaPath)) {
        if(tester.test(line)) results.push(line.split('\t').reduce((p,c,i) => ({...p, [header[i]] : separator.test(c) ? c.split('|') : c}), {}))
    }

    const subTerms = results.map(e => e?.GO_Term)

    const cacheTerms = await getGOterms(subTerms, oboPath)
    for (let index = 0; index < results.length; index++) {
        const goTerm = results[index]?.GO_Term
        results[index]['GO_Term'] = cacheTerms.filter(e => e.id === goTerm)[0]
    }
    return results
}

const getGOterms = async (
    terms: string[],
    oboPath: string
) => {
    const testerList = terms.map(e => new RegExp('id: ' + e))
    let delim = false
    const results = []
    let result = {} as {[key:string]: any}
    for await (const line of line$(oboPath)) {
        if (testerList.some(rx => rx.test(line))) delim = true
        if (line === '' && delim) {
            delim = false
            results.push(result)
            result = {}
        }
        if (delim) result[line.split(': ')[0]] = line.split(': ')[1]
    }
    return results
}

export { getSymbol, getGOterms }