index.js 8.2 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183
  1. "use strict";
  2. var __awaiter = (this && this.__awaiter) || function (thisArg, _arguments, P, generator) {
  3. function adopt(value) { return value instanceof P ? value : new P(function (resolve) { resolve(value); }); }
  4. return new (P || (P = Promise))(function (resolve, reject) {
  5. function fulfilled(value) { try { step(generator.next(value)); } catch (e) { reject(e); } }
  6. function rejected(value) { try { step(generator["throw"](value)); } catch (e) { reject(e); } }
  7. function step(result) { result.done ? resolve(result.value) : adopt(result.value).then(fulfilled, rejected); }
  8. step((generator = generator.apply(thisArg, _arguments || [])).next());
  9. });
  10. };
  11. var __asyncValues = (this && this.__asyncValues) || function (o) {
  12. if (!Symbol.asyncIterator) throw new TypeError("Symbol.asyncIterator is not defined.");
  13. var m = o[Symbol.asyncIterator], i;
  14. 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);
  15. 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); }); }; }
  16. function settle(resolve, reject, d, v) { Promise.resolve(v).then(function(v) { resolve({ value: v, done: d }); }, reject); }
  17. };
  18. var __importDefault = (this && this.__importDefault) || function (mod) {
  19. return (mod && mod.__esModule) ? mod : { "default": mod };
  20. };
  21. Object.defineProperty(exports, "__esModule", { value: true });
  22. exports.getAllSymbols = exports.getN = exports.getGOterms = exports.getSymbol = void 0;
  23. const fs_1 = __importDefault(require("fs"));
  24. const readline_1 = __importDefault(require("readline"));
  25. const stats_1 = require("@stdlib/stats");
  26. const line$ = (path) => readline_1.default.createInterface({
  27. input: fs_1.default.createReadStream(path),
  28. crlfDelay: Infinity
  29. });
  30. // http://geneontology.org/docs/guide-go-evidence-codes/
  31. // http://wiki.geneontology.org/index.php/Category:Evidence_Codes
  32. // http://current.geneontology.org/ontology/go-basic.obo
  33. // http://current.geneontology.org/annotations/goa_human.gaf.gz
  34. const getSymbol = (symbol, goaPath, oboPath) => __awaiter(void 0, void 0, void 0, function* () {
  35. var e_1, _a;
  36. var _b;
  37. const header = [
  38. 'database', 'ID', 'Symbol', 'Qualifier',
  39. 'GO_Term', 'Evidence', 'Evidence_Code',
  40. 'With', 'From', 'Name', 'Alternative_symbols',
  41. 'Class', 'Taxon', 'Date', 'Origin'
  42. ];
  43. const tester = new RegExp('\t' + symbol + '\t');
  44. const separator = new RegExp('\\|');
  45. const results = [];
  46. try {
  47. for (var _c = __asyncValues(line$(goaPath)), _d; _d = yield _c.next(), !_d.done;) {
  48. const line = _d.value;
  49. if (tester.test(line))
  50. results.push(line.split('\t').filter((e) => e).reduce((p, c, i) => (Object.assign(Object.assign({}, p), { [header[i]]: separator.test(c) ? c.split('|') : c })), {}));
  51. }
  52. }
  53. catch (e_1_1) { e_1 = { error: e_1_1 }; }
  54. finally {
  55. try {
  56. if (_d && !_d.done && (_a = _c.return)) yield _a.call(_c);
  57. }
  58. finally { if (e_1) throw e_1.error; }
  59. }
  60. const subTerms = results.map(e => e === null || e === void 0 ? void 0 : e.GO_Term);
  61. const cacheTerms = yield getGOterms(subTerms, oboPath);
  62. for (let index = 0; index < results.length; index++) {
  63. const goTerm = (_b = results[index]) === null || _b === void 0 ? void 0 : _b.GO_Term;
  64. results[index]['GO_Term'] = cacheTerms.filter(e => e.id === goTerm)[0];
  65. }
  66. return results;
  67. });
  68. exports.getSymbol = getSymbol;
  69. const getGOterms = (terms, oboPath) => __awaiter(void 0, void 0, void 0, function* () {
  70. var e_2, _e;
  71. terms = Array.isArray(terms) ? terms : [terms];
  72. const testerList = terms.map(e => new RegExp('id: ' + e));
  73. let delim = false;
  74. const results = [];
  75. let result = {};
  76. try {
  77. for (var _f = __asyncValues(line$(oboPath)), _g; _g = yield _f.next(), !_g.done;) {
  78. const line = _g.value;
  79. if (testerList.some(rx => rx.test(line)))
  80. delim = true;
  81. if (line === '' && delim) {
  82. delim = false;
  83. results.push(result);
  84. result = {};
  85. }
  86. if (delim)
  87. result[line.split(': ')[0]] = line.split(': ')[1];
  88. }
  89. }
  90. catch (e_2_1) { e_2 = { error: e_2_1 }; }
  91. finally {
  92. try {
  93. if (_g && !_g.done && (_e = _f.return)) yield _e.call(_f);
  94. }
  95. finally { if (e_2) throw e_2.error; }
  96. }
  97. return results;
  98. });
  99. exports.getGOterms = getGOterms;
  100. const getN = (terms, qualifiers, goaPath, oboPath) => __awaiter(void 0, void 0, void 0, function* () {
  101. var e_3, _h;
  102. terms = Array.isArray(terms) ? terms : [terms];
  103. qualifiers = Array.isArray(qualifiers) ? qualifiers : [qualifiers];
  104. let qualifs = {};
  105. let termsList = [];
  106. for (const term of terms) {
  107. termsList.push(Object.assign(Object.assign({ nTotal: 0, genes: [] }, (yield getGOterms(term, oboPath))[0]), { test: new RegExp('\t' + term + '\t') }));
  108. }
  109. qualifiers.map((e) => qualifs[e] = {});
  110. try {
  111. for (var _j = __asyncValues(line$(goaPath)), _k; _k = yield _j.next(), !_k.done;) {
  112. const line = _k.value;
  113. const t = termsList.filter((rx) => rx.test.test(line));
  114. if (t.length > 0) {
  115. termsList = termsList.map((rx) => {
  116. if (rx.test.test(line)) {
  117. const genes = [...new Set([...(rx === null || rx === void 0 ? void 0 : rx.genes), line.split('\t')[2]])];
  118. return Object.assign(Object.assign({}, rx), { nTotal: genes.length, genes });
  119. }
  120. else {
  121. return rx;
  122. }
  123. });
  124. }
  125. }
  126. }
  127. catch (e_3_1) { e_3 = { error: e_3_1 }; }
  128. finally {
  129. try {
  130. if (_k && !_k.done && (_h = _j.return)) yield _h.call(_j);
  131. }
  132. finally { if (e_3) throw e_3.error; }
  133. }
  134. return termsList.map((e) => {
  135. delete e['test'];
  136. return e;
  137. });
  138. });
  139. exports.getN = getN;
  140. const headerGOA = [
  141. 'database', 'ID', 'Symbol', 'Qualifier',
  142. 'GO_Term', 'Evidence', 'Evidence_Code',
  143. 'With', 'From', 'Name', 'Alternative_symbols',
  144. 'Class', 'Taxon', 'Date', 'Origin'
  145. ];
  146. const getAllSymbols = (symbols, qualifier, goaPath, oboPath) => __awaiter(void 0, void 0, void 0, function* () {
  147. var e_4, _l;
  148. symbols = Array.isArray(symbols) ? symbols : [symbols];
  149. const testerList = [...new Set(symbols)].map(e => new RegExp('\t' + e + '\t' + qualifier + '\t'));
  150. const allGoa = [];
  151. const allGenes = {};
  152. try {
  153. for (var _m = __asyncValues(line$(goaPath)), _o; _o = yield _m.next(), !_o.done;) {
  154. const line = _o.value;
  155. if (testerList.some(rx => rx.test(line))) {
  156. allGoa.push(line.split('\t').
  157. reduce((p, c, i) => (Object.assign(Object.assign({}, p), { [headerGOA[i]]: c })), {}));
  158. }
  159. allGenes[line.split('\t')[2]] = {};
  160. }
  161. }
  162. catch (e_4_1) { e_4 = { error: e_4_1 }; }
  163. finally {
  164. try {
  165. if (_o && !_o.done && (_l = _m.return)) yield _l.call(_m);
  166. }
  167. finally { if (e_4) throw e_4.error; }
  168. }
  169. const nAllGenes = Object.keys(allGenes).length;
  170. console.log(nAllGenes);
  171. const allIDs = yield getN([...new Set(allGoa.map((e) => e.GO_Term))], qualifier, goaPath, oboPath);
  172. return allIDs.map((e) => {
  173. const observedGenes = symbols.filter((ee) => e.genes.includes(ee));
  174. // const observedGenesNotIn = symbols.filter((ee:string)=> !e.genes.includes(ee))
  175. // const allIn = Object.keys(allGenes).filter((ee:any)=>)
  176. const diff = e.genes.length - observedGenes.length;
  177. return Object.assign({ n: observedGenes.length, prop: observedGenes.length / e.genes.length, chi2test: (0, stats_1.chi2test)([[observedGenes.length,
  178. diff < 0 ? 0 : diff],
  179. [symbols.length,
  180. nAllGenes]]).pValue, observedGenes }, e);
  181. }).sort((a, b) => b.chi2test - a.chi2test);
  182. });
  183. exports.getAllSymbols = getAllSymbols;