gbffParser/index.ts

// wget https://ftp.ncbi.nlm.nih.gov/genomes/all/annotation_releases/9606/109.20211119/GCF_000001405.39_GRCh38.p13/GCF_000001405.39_GRCh38.p13_feature_table.txt.gz
// wget https://ftp.ncbi.nlm.nih.gov/genomes/all/annotation_releases/9606/109.20211119/GCF_000001405.39_GRCh38.p13/GCF_000001405.39_GRCh38.p13_assembly_structure/Primary_Assembly/assembled_chromosomes/chr2acc
// wget https://ftp.ncbi.nlm.nih.gov/genomes/all/annotation_releases/9606/109.20211119/GCF_000001405.39_GRCh38.p13/GCF_000001405.39_GRCh38.p13_genomic.gbff.gz
// wget ftp://ftp.ncbi.nlm.nih.gov/refseq/H_sapiens/RefSeqGene/LRG_RefSeqGene
// wget ftp://ftp.ncbi.nlm.nih.gov/refseq/H_sapiens/RefSeqGene/refseqgene.1.genomic.gbff.gz
// wget ftp://ftp.ncbi.nlm.nih.gov/refseq/H_sapiens/biological_region/human.biological_region.gbff.gz
// wget ftp://ftp.ncbi.nlm.nih.gov/refseq/H_sapiens/mRNA_Prot/human.[1-10].rna.gbff.gz

import fs from 'fs'
import os from 'os'
import path from 'path'
import { spawn } from 'child_process'
import readline from 'readline'
import { Buffer } from 'buffer'
import genbankParser from 'genbank-parser'
import { writeSequence } from 'aligner'
import jsonata from 'jsonata'

const async_exec = (prog: string, args: string[], onData: Function) => {
    return new Promise((resolve, reject) => {
        const child = spawn(prog, args, {shell: true})

        child.stdout.on('data', data => onData(data.toString().trim()))
        child.stderr.on('data', data => onData(data.toString().trim()))

        child.on('error', err => reject(err))
        child.on('exit', code => resolve(code))
    })
}

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

const readOffset = (path: string, from:number, to:number) => {
    return new Promise<string>(async (resolve, reject) => {
        const size = to - from
        const buffer = Buffer.alloc(size)
        let filehandle = null
        try {
            filehandle = await fs.promises.open(path, 'r+');
            await filehandle.read(buffer, 0, buffer.length, from)
        } finally {
            if (filehandle) {
                await filehandle.close()
                resolve(buffer.toString())
            }
        }
    })
}

/*
 * strings -a -t d human.1.rna.gbff | grep LOCUS | awk '{print $1"\t"$3}' > human.1.rna.gbff.index
 * 
 */
const makeGbffIndex = async (filePath: string, indexPath?: string) => {
    interface entry {
        filePath: string;
        value   : string;
        from    : number;
        to     ?: number;
    }
    indexPath = indexPath || filePath + '.jsi'
    let entries = [] as entry[]
    let lineN   = 0
    let byteAcc = 0
    for await (const line of line$(filePath)) {
        if(line.match(/^LOCUS/)) {
            entries.push({
                filePath,
                value   : line.split(/\s+/)[1],
                from    : byteAcc
            })
            if(lineN !== 0) {
                entries[entries.length - 2]["to"] = byteAcc 
                await fs.promises.appendFile(indexPath, [
                    entries[entries.length - 2]["value"],
                    entries[entries.length - 2]["from"],
                    entries[entries.length - 2]["to"]].join('\t') + '\n')
                entries = entries.splice(1)
            }
        }
        byteAcc += (line.length + 1)
        lineN++
    }
    entries[entries.length - 1]["to"] = byteAcc
    
    await fs.promises.appendFile(indexPath, [
        entries[entries.length - 1]["value"],
        entries[entries.length - 1]["from"],
        entries[entries.length - 1]["to"]].join('\t'))
    return entries
}

const getOffset = async (indexPath: string, acc: string) => {
    let res
    for await (const line of line$(indexPath)) {
        const tmp = line.split('\t')
        if (tmp[0] === acc) {
            res = [indexPath.split('.jsi')[0], tmp[1], tmp[2]]
            break
        }
    }
    return res
}

const getOffsets = async (indexPath: string | string[], accessions: RegExp) => {
    let res: string[][] = []
    const indexPaths = Array.isArray(indexPath) ? indexPath : [indexPath]
    for (const iP of indexPaths) {
        for await (const line of line$(iP)) {
            const tmp = line.split('\t')
            if (accessions.test(tmp[0])) {
                res.push([iP.split('.jsi')[0], tmp[1], tmp[2], tmp[0]])
            }
        }
    }
    return res
}

const getData =  async (dbPath: string | string[], accessionRegex: RegExp, query?: string) => {
    dbPath = Array.isArray(dbPath) ? dbPath : [dbPath]
    const results = []
    const allOffsets = await getOffsets(dbPath.map(e => e + '.jsi'), accessionRegex)
    for (const offset of allOffsets) {
        const txt = await readOffset(offset[0], Number(offset[1]), Number(offset[2]))
        const json = genbankParser(txt)[0]
        query ? results.push(jsonata(query).evaluate(json)) : results.push(json)
    }
    return results
}

const getFromAcc = async (accession: string, dbPath: string | string[], indexPath?: string | string[]) => {
    dbPath = Array.isArray(dbPath) ? dbPath : [dbPath]
    if (!indexPath) {
        indexPath = await getJSI(dbPath)
    } else {
        indexPath = Array.isArray(indexPath) ? indexPath : [indexPath]
        if (indexPath.length !== dbPath.length) throw 'Error'
    }

    for (const iP of indexPath) {
        const [filePath, from, to] = await getOffset(iP, accession) || [undefined,undefined,undefined]
        if (filePath) {
            const txt = await readOffset(filePath, Number(from), Number(to))
            return genbankParser(txt)[0]
        }
    }
    return undefined
}

interface selector {[key: string]: string}
const getPos = async (selector: selector, tablePath: string, headerTablePath:string[]) => {
    const results:any = [];
    (await fs.promises.readFile(tablePath)).toString().split('\n')
    .map((line:string) => {
        const lineObj = line.split('\t').reduce((p,c,i) => ({...p, [headerTablePath[i]]: isFinite(Number(c)) ? Number(c) : c}), {} as any)
        if(lineObj[Object.keys(selector)[0]] === selector[Object.keys(selector)[0]]) results.push(lineObj)
    })
    return results
}

const getIndex = async (symbol:string, LRGPath:string) => {
    interface geneIndex  {
        GeneID        : string;
        GeneAcc       : string;
        TranscriptsAcc: string[];
        ProteinAcc    : string[]
    }
    return (await fs.promises.readFile(LRGPath)).toString()
    .split('\n')
    .filter((line:string) => line.match(new RegExp(symbol, 'g')))
    .reduce((p:any,c:any) => {
        const [TaxID, GeneID, GeneName, GeneAcc, TranscriptsAcc, ProteinAcc, _] = c.split('\t').filter((e:any)=>e!=='')
        return {GeneID, GeneAcc, TranscriptsAcc: [...(new Set([...p.TranscriptsAcc, TranscriptsAcc]))], ProteinAcc: [...(new Set([...p.ProteinAcc, ProteinAcc]))]}
    }, {GeneID: '', GeneAcc: '', TranscriptsAcc: [], ProteinAcc: []} as geneIndex)
}

const getSymbol = async (
    symbol:string,
    LRGPath:string, // wget ftp://ftp.ncbi.nlm.nih.gov/refseq/H_sapiens/RefSeqGene/LRG_RefSeqGene
    tablePath:string, // wget https://ftp.ncbi.nlm.nih.gov/genomes/all/annotation_releases/9606/109.20211119/GCF_000001405.39_GRCh38.p13/GCF_000001405.39_GRCh38.p13_feature_table.txt.gz
    // regionDBPath: string | string[], // wget ftp://ftp.ncbi.nlm.nih.gov/refseq/H_sapiens/biological_region/human.biological_region.gbff.gz
    geneDBPath: string | string[], // wget ftp://ftp.ncbi.nlm.nih.gov/refseq/H_sapiens/RefSeqGene/refseqgene.[1-7].genomic.gbff.gz
    rnaDBPath: string | string[] // wget ftp://ftp.ncbi.nlm.nih.gov/refseq/H_sapiens/mRNA_Prot/human.[1-10].rna.gbff.gz
) => {
    const geneIndex = await getIndex(symbol, LRGPath)

    // const regionData = await getFromAcc(geneIndex.GeneAcc.split('.')[0], regionDBPath)
    // if (regionData) await fs.promises.writeFile('test-region.json', JSON.stringify(regionData, null, 4))

    const headerTablePath = [
        'feature', 'class', 'assembly', 'assembly_unit', 'seq_type', 'chromosome',
        'genomic_accession', 'start', 'end', 'strand', 'product_accession', 'non_redundant_refseq',
        'related_accession', 'name', 'symbol', 'GeneID', 'locus_tag', 'feature_interval_length',
        'product_length','attributes'
    ]
    const allFeatures = await getPos({symbol}, tablePath, headerTablePath)

    for (let index = 0; index < allFeatures.length; index++) {
        const {feature, product_accession} = allFeatures[index]
        let tmp
        switch (feature) {
            case 'gene':
                allFeatures[index].product_accession = geneIndex.GeneAcc
                tmp = await getFromAcc(geneIndex.GeneAcc.split('.')[0], geneDBPath)
                // await fs.promises.writeFile('test/test-gene.json', JSON.stringify(tmp, null, 4))
                allFeatures[index].data = tmp
                break;
            case 'mRNA':
                tmp = await getFromAcc(product_accession.split('.')[0], rnaDBPath)
                // await fs.promises.writeFile('test/test-rna-'+index+'.json', JSON.stringify(tmp, null, 4))
                allFeatures[index].data = tmp
                break;
            default:
                break;
        }
        
    }
    return allFeatures
}

const getJSI = async (dbPath: string | string[]) => {
    dbPath = Array.isArray(dbPath) ? dbPath : [dbPath]
    const indexPath: string[] = []
    for (const p of dbPath) {
        const iP = p + '.jsi'
        if (!fs.existsSync(iP)) {
            console.log('Writing index: ' + iP);
            await makeGbffIndex(p)
        }
        indexPath.push(iP)
    }
    return indexPath
}

// Todo: add progress
const makeRefSeqFromReg = async (
    dbPath: string | string[], reg: RegExp, distFile:string, limit?: number
) => {
    dbPath = Array.isArray(dbPath) ? dbPath : [dbPath]
    const jsiFiles = await getJSI(dbPath)
    const tmpDir = path.join(os.tmpdir(), 'parser-' + Math.random())
    await fs.promises.mkdir(tmpDir)
    const createdFiles: string[] = []
    let counter = 0

    for (const jsiFile of jsiFiles) {
        console.log('reading ' + jsiFile)
        for await (const line of line$(jsiFile)) {
            if(line.match(reg)) {
                const [accession, from, to] = line.split('\t')
                const res = await getFromAcc(accession, jsiFile.split('.jsi')[0])
                if (res?.sequence) {
                    try {
                        const file = path.join(tmpDir, res?.version || res.accession + '.fa')
                        if (!createdFiles.includes(file)) {
                            if (createdFiles.length === 0) if (fs.existsSync(distFile)) await fs.promises.rm(distFile)
                            await writeSequence(res?.version || res.accession, res?.sequence, file)
                            createdFiles.push(file)
                            const tmp = await fs.promises.readFile(file)
                            await fs.promises.appendFile(distFile, tmp.toString() + '\n')
                            await fs.promises.rm(file)
                            counter++
                            if (counter%100 === 0) console.log('Already ' + counter + ' sequence parsed')
                        }
                    } catch (error) {
                        console.log(error)
                    }
                }
            }
            if (limit) if (counter === limit) break
        }
        if (limit) if (counter === limit) break
    }
    console.log(createdFiles.length + ' sequences were extracted')
    
    await fs.promises.rm(tmpDir, {recursive: true})
    await async_exec('bwa', ['index', distFile], () => console.log)
}

export { getFromAcc, getSymbol, makeRefSeqFromReg, getOffsets, getData }