| 123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230 |
- "use strict";
- // https://ftp.uniprot.org/pub/databases/uniprot/current_release/knowledgebase/taxonomic_divisions/uniprot_sprot_human.xml.gz
- var __awaiter = (this && this.__awaiter) || function (thisArg, _arguments, P, generator) {
- function adopt(value) { return value instanceof P ? value : new P(function (resolve) { resolve(value); }); }
- return new (P || (P = Promise))(function (resolve, reject) {
- function fulfilled(value) { try { step(generator.next(value)); } catch (e) { reject(e); } }
- function rejected(value) { try { step(generator["throw"](value)); } catch (e) { reject(e); } }
- function step(result) { result.done ? resolve(result.value) : adopt(result.value).then(fulfilled, rejected); }
- step((generator = generator.apply(thisArg, _arguments || [])).next());
- });
- };
- var __asyncValues = (this && this.__asyncValues) || function (o) {
- if (!Symbol.asyncIterator) throw new TypeError("Symbol.asyncIterator is not defined.");
- var m = o[Symbol.asyncIterator], i;
- return m ? m.call(o) : (o = typeof __values === "function" ? __values(o) : o[Symbol.iterator](), i = {}, verb("next"), verb("throw"), verb("return"), i[Symbol.asyncIterator] = function () { return this; }, i);
- function verb(n) { i[n] = o[n] && function (v) { return new Promise(function (resolve, reject) { v = o[n](v), settle(resolve, reject, v.done, v.value); }); }; }
- function settle(resolve, reject, d, v) { Promise.resolve(v).then(function(v) { resolve({ value: v, done: d }); }, reject); }
- };
- var __importDefault = (this && this.__importDefault) || function (mod) {
- return (mod && mod.__esModule) ? mod : { "default": mod };
- };
- Object.defineProperty(exports, "__esModule", { value: true });
- exports.getInteractionsFromEntry = exports.getEntryFromGeneName = exports.getEnrty = exports.readOffset = exports.makeIndex = void 0;
- const fast_xml_parser_1 = require("fast-xml-parser");
- const fs_1 = __importDefault(require("fs"));
- const readline_1 = __importDefault(require("readline"));
- const line$ = (path) => readline_1.default.createInterface({
- input: fs_1.default.createReadStream(path),
- crlfDelay: Infinity
- });
- const makeIndex = (filePath, indexPath) => __awaiter(void 0, void 0, void 0, function* () {
- var e_1, _a;
- indexPath = indexPath || filePath + '.jsi';
- let byteAcc = 0;
- const fromSel = new RegExp("^<entry");
- const toSel = new RegExp("^</entry>");
- const valSel = new RegExp('<accession>');
- let tmp = { values: [] };
- try {
- for (var _b = __asyncValues(line$(filePath)), _c; _c = yield _b.next(), !_c.done;) {
- const line = _c.value;
- if (fromSel.test(line))
- tmp['from'] = byteAcc;
- byteAcc += (line.length + 1);
- if (valSel.test(line))
- tmp['values'].push(line.match("<accession>(.*?)</accession>")[1]); // 'uck
- if (toSel.test(line)) {
- yield fs_1.default.promises.appendFile(indexPath, tmp.values.join(';') + '\t' + tmp.from + '\t' + byteAcc + '\n');
- tmp = { values: [] };
- }
- }
- }
- catch (e_1_1) { e_1 = { error: e_1_1 }; }
- finally {
- try {
- if (_c && !_c.done && (_a = _b.return)) yield _a.call(_b);
- }
- finally { if (e_1) throw e_1.error; }
- }
- });
- exports.makeIndex = makeIndex;
- const readOffset = (path, from, to) => {
- return new Promise((resolve, reject) => __awaiter(void 0, void 0, void 0, function* () {
- const size = to - from;
- const buffer = Buffer.alloc(size);
- let filehandle;
- try {
- filehandle = yield fs_1.default.promises.open(path, 'r+');
- yield filehandle.read(buffer, 0, buffer.length, from);
- }
- finally {
- if (filehandle) {
- yield filehandle.close();
- resolve(buffer.toString());
- }
- }
- }));
- };
- exports.readOffset = readOffset;
- const getEntryOffset = (dbPath, accession) => __awaiter(void 0, void 0, void 0, function* () {
- var e_2, _d;
- const indexPath = dbPath + '.jsi';
- if (!fs_1.default.existsSync(indexPath))
- yield makeIndex(dbPath);
- const lineSel = new RegExp(accession);
- try {
- for (var _e = __asyncValues(line$(indexPath)), _f; _f = yield _e.next(), !_f.done;) {
- const line = _f.value;
- if (lineSel.test(line))
- return [Number(line.split('\t')[1]), Number(line.split('\t')[2])];
- }
- }
- catch (e_2_1) { e_2 = { error: e_2_1 }; }
- finally {
- try {
- if (_f && !_f.done && (_d = _e.return)) yield _d.call(_e);
- }
- finally { if (e_2) throw e_2.error; }
- }
- return [0, 0];
- });
- const getEntry = (dbPath, accession) => __awaiter(void 0, void 0, void 0, function* () {
- const parser = new fast_xml_parser_1.XMLParser({
- ignoreAttributes: false,
- alwaysCreateTextNode: false,
- attributeNamePrefix: "",
- textNodeName: "value",
- allowBooleanAttributes: true,
- });
- const offsets = yield getEntryOffset(dbPath, accession);
- return parser.parse(yield readOffset(dbPath, offsets[0], offsets[1]));
- });
- exports.getEnrty = getEntry;
- const getEntryFromGeneName = (idmappingPath, dbPath, geneName) => __awaiter(void 0, void 0, void 0, function* () {
- const accessions = yield getAccessFromGene(idmappingPath, geneName);
- return yield getEntry(dbPath, accessions[0]); // seems to be always the first with entry
- });
- exports.getEntryFromGeneName = getEntryFromGeneName;
- const getAccessFromGene = (idmappingPath, geneName) => __awaiter(void 0, void 0, void 0, function* () {
- var e_3, _g;
- const sel = new RegExp('Gene_Name\t' + geneName);
- let accessions = [];
- try {
- for (var _h = __asyncValues(line$(idmappingPath)), _j; _j = yield _h.next(), !_j.done;) {
- const line = _j.value;
- if (sel.test(line))
- accessions.push(line.split('\t')[0]);
- }
- }
- catch (e_3_1) { e_3 = { error: e_3_1 }; }
- finally {
- try {
- if (_j && !_j.done && (_g = _h.return)) yield _g.call(_h);
- }
- finally { if (e_3) throw e_3.error; }
- }
- return accessions;
- });
- const getInteractionsFromEntry = (json) => __awaiter(void 0, void 0, void 0, function* () {
- var _k, _l, _m, _o, _p, _q, _r, _s;
- const blaskList = ['DNA', 'PHOSPHOSERINE', 'MOTIFS', 'INFECTION', 'PROTEIN', 'PROTEINS', 'GAMMA-SECRETASE', 'CALCIUM',
- 'MICROBIAL', 'VIRUS', 'HEPATITIS', 'HERPES', 'SIMPLEX', 'RELATED', 'AND', 'CLATHRIN', 'WORTMANNIN',
- 'NUCLEOSOME', 'undefined', 'INTEGRINS', 'UBIQUITIN', 'MAGNESIUM'];
- const uniprotIDs = Array.isArray(json.entry.accession) ? json.entry.accession : [json.entry.accession];
- // geneName
- const gnTT = Array.isArray(json.entry.gene) ? json.entry.gene[0] : json.entry.gene;
- let geneName = '';
- if (gnTT === null || gnTT === void 0 ? void 0 : gnTT.name) {
- const gnT = Array.isArray(gnTT.name) ? gnTT.name : [gnTT.name];
- geneName = gnT.filter((e) => e.type === 'primary').map((e) => e.value)[0];
- }
- else if ((_l = (_k = json.entry) === null || _k === void 0 ? void 0 : _k.protein) === null || _l === void 0 ? void 0 : _l.recommendedName) {
- geneName = Array.isArray((_o = (_m = json.entry) === null || _m === void 0 ? void 0 : _m.protein) === null || _o === void 0 ? void 0 : _o.recommendedName) ? (_q = (_p = json.entry) === null || _p === void 0 ? void 0 : _p.protein) === null || _q === void 0 ? void 0 : _q.recommendedName[0] : (_s = (_r = json.entry) === null || _r === void 0 ? void 0 : _r.protein) === null || _s === void 0 ? void 0 : _s.recommendedName;
- }
- // Interactants
- const jecT = Array.isArray(json.entry.comment) ? json.entry.comment : [json.entry.comment];
- const interactants = jecT
- .filter((e) => (e === null || e === void 0 ? void 0 : e.type) === 'interaction')
- .flatMap((e) => ({
- type: 'interactant',
- fromProductId: e.interactant[0].id,
- toProductId: e.interactant[1].id,
- to: e.interactant[1].label,
- nExperiments: Number(e.experiments)
- }));
- const regExp = new RegExp('INTERACTION WITH |Interacts with |complex with ', 'i');
- const geneRegExp = new RegExp(/[A-Z]{1}[A-Z|0-9]{2,}$/);
- // uniprot_comment_text_value
- const commentInteractsWith = jecT
- .filter((e) => { var _a; return (_a = e === null || e === void 0 ? void 0 : e.text) === null || _a === void 0 ? void 0 : _a.value; })
- .filter((e) => regExp.test(e.text.value))
- .map((e) => ({
- to: e.text.value
- .split(/\.|;/)
- .flatMap((ee) => ee.replace(/ *\([^)]*\) */g, ' '))
- .filter((ee) => regExp.test(ee))
- .flatMap((ee) => ee.trim().split(regExp))
- .flatMap((ee) => ee.split(/,| and | /))
- .filter((_) => _)
- .filter((ee) => geneRegExp.test(ee))
- .filter((ee) => !blaskList.includes(ee) && ee !== geneName)
- .map((ee) => ee.trim()),
- text: e.text.value,
- //evidences: e.text.evidence.split(' ')//.map((ee:string)=> json.entry.reference.filter((eee:any)=> eee.key === ee)) // Doesnt work with ref key
- }))
- .flatMap((e) => e.to.flatMap((ee) => ({
- type: 'uniprot_comment_text_value',
- to: ee,
- text: e.text
- })));
- // uniprot_reference_scope
- const jerT = Array.isArray(json.entry.reference) ? json.entry.reference : [json.entry.reference];
- const referenceInteract = jerT
- .map((e) => (Object.assign(Object.assign({}, e), { scope: Array.isArray(e.scope) ? e.scope : [e.scope] })))
- .filter((e) => regExp.test(e.scope.join('')))
- .map((e) => (Object.assign({ to: e.scope
- //.split(/\.|;/)
- .flatMap((ee) => regExp.test(ee) ? [ee] : [])
- .filter((_) => _)
- .flatMap((ee) => ee.replace(/ *\([^)]*\) */g, ' '))
- .filter((ee) => regExp.test(ee))
- .flatMap((ee) => ee.trim().split(regExp)[1])
- .flatMap((ee) => ee.split(/,| and | /i))
- .filter((_) => _)
- .filter((ee) => geneRegExp.test(ee))
- .filter((ee) => !blaskList.includes(ee) && ee !== geneName)
- .map((ee) => ee.trim()) }, e)))
- .flatMap((e) => e.to.flatMap((ee) => ({
- type: 'reference_scope',
- to: ee,
- scope: e.scope,
- citation: e.citation
- })));
- // Group
- const byTo = {};
- [...interactants, ...referenceInteract, ...commentInteractsWith]
- .map((e) => byTo[e.to] = byTo[e.to] ? Object.assign(Object.assign({}, e), byTo[e.to]) : Object.assign({}, e));
- const results = Object.keys(byTo).map((e) => {
- var _a;
- (_a = byTo[e]) === null || _a === void 0 ? true : delete _a.to;
- return {
- from: geneName,
- to: e,
- data: byTo[e]
- };
- })
- .filter((e) => !blaskList.includes(e.to) && e.to !== geneName);
- return results;
- });
- exports.getInteractionsFromEntry = getInteractionsFromEntry;
|