index.js 14 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259
  1. "use strict";
  2. var __assign = (this && this.__assign) || function () {
  3. __assign = Object.assign || function(t) {
  4. for (var s, i = 1, n = arguments.length; i < n; i++) {
  5. s = arguments[i];
  6. for (var p in s) if (Object.prototype.hasOwnProperty.call(s, p))
  7. t[p] = s[p];
  8. }
  9. return t;
  10. };
  11. return __assign.apply(this, arguments);
  12. };
  13. var __awaiter = (this && this.__awaiter) || function (thisArg, _arguments, P, generator) {
  14. function adopt(value) { return value instanceof P ? value : new P(function (resolve) { resolve(value); }); }
  15. return new (P || (P = Promise))(function (resolve, reject) {
  16. function fulfilled(value) { try { step(generator.next(value)); } catch (e) { reject(e); } }
  17. function rejected(value) { try { step(generator["throw"](value)); } catch (e) { reject(e); } }
  18. function step(result) { result.done ? resolve(result.value) : adopt(result.value).then(fulfilled, rejected); }
  19. step((generator = generator.apply(thisArg, _arguments || [])).next());
  20. });
  21. };
  22. var __generator = (this && this.__generator) || function (thisArg, body) {
  23. var _ = { label: 0, sent: function() { if (t[0] & 1) throw t[1]; return t[1]; }, trys: [], ops: [] }, f, y, t, g;
  24. return g = { next: verb(0), "throw": verb(1), "return": verb(2) }, typeof Symbol === "function" && (g[Symbol.iterator] = function() { return this; }), g;
  25. function verb(n) { return function (v) { return step([n, v]); }; }
  26. function step(op) {
  27. if (f) throw new TypeError("Generator is already executing.");
  28. while (_) try {
  29. if (f = 1, y && (t = op[0] & 2 ? y["return"] : op[0] ? y["throw"] || ((t = y["return"]) && t.call(y), 0) : y.next) && !(t = t.call(y, op[1])).done) return t;
  30. if (y = 0, t) op = [op[0] & 2, t.value];
  31. switch (op[0]) {
  32. case 0: case 1: t = op; break;
  33. case 4: _.label++; return { value: op[1], done: false };
  34. case 5: _.label++; y = op[1]; op = [0]; continue;
  35. case 7: op = _.ops.pop(); _.trys.pop(); continue;
  36. default:
  37. if (!(t = _.trys, t = t.length > 0 && t[t.length - 1]) && (op[0] === 6 || op[0] === 2)) { _ = 0; continue; }
  38. if (op[0] === 3 && (!t || (op[1] > t[0] && op[1] < t[3]))) { _.label = op[1]; break; }
  39. if (op[0] === 6 && _.label < t[1]) { _.label = t[1]; t = op; break; }
  40. if (t && _.label < t[2]) { _.label = t[2]; _.ops.push(op); break; }
  41. if (t[2]) _.ops.pop();
  42. _.trys.pop(); continue;
  43. }
  44. op = body.call(thisArg, _);
  45. } catch (e) { op = [6, e]; y = 0; } finally { f = t = 0; }
  46. if (op[0] & 5) throw op[1]; return { value: op[0] ? op[1] : void 0, done: true };
  47. }
  48. };
  49. var __spreadArray = (this && this.__spreadArray) || function (to, from, pack) {
  50. if (pack || arguments.length === 2) for (var i = 0, l = from.length, ar; i < l; i++) {
  51. if (ar || !(i in from)) {
  52. if (!ar) ar = Array.prototype.slice.call(from, 0, i);
  53. ar[i] = from[i];
  54. }
  55. }
  56. return to.concat(ar || Array.prototype.slice.call(from));
  57. };
  58. var __importDefault = (this && this.__importDefault) || function (mod) {
  59. return (mod && mod.__esModule) ? mod : { "default": mod };
  60. };
  61. Object.defineProperty(exports, "__esModule", { value: true });
  62. exports.makeReference = exports.writeSequence = exports.asyncBwaMem = void 0;
  63. var child_process_1 = require("child_process");
  64. var os_1 = require("os");
  65. var fs_1 = __importDefault(require("fs"));
  66. var path_1 = __importDefault(require("path"));
  67. var async_exec = function (prog, args, onData) {
  68. return new Promise(function (resolve, reject) {
  69. var child = (0, child_process_1.spawn)(prog, args, { shell: true });
  70. child.stdout.on('data', function (data) { return onData(data.toString().trim()); });
  71. child.stderr.on('data', function (data) { return onData(data.toString().trim()); });
  72. child.on('error', function (err) { return reject(err); });
  73. child.on('exit', function (code) { return resolve(code); });
  74. });
  75. };
  76. var invReplace = function (regex, string, by) {
  77. if (by === void 0) { by = '_'; }
  78. return string.split('').map(function (letter) { return letter.match(regex) ? letter : by; }).join('');
  79. };
  80. var writeSequence = function (sequenceName, sequence, filePath, lineN) {
  81. if (lineN === void 0) { lineN = 80; }
  82. return __awaiter(void 0, void 0, void 0, function () {
  83. return __generator(this, function (_a) {
  84. return [2 /*return*/, new Promise(function (resolve, reject) { return __awaiter(void 0, void 0, void 0, function () {
  85. var r, regex_sam_restriction, nSeqName, error_1;
  86. var _a;
  87. return __generator(this, function (_b) {
  88. switch (_b.label) {
  89. case 0:
  90. _b.trys.push([0, 2, , 3]);
  91. r = new RegExp(".{1," + lineN + "}", "g");
  92. regex_sam_restriction = /[>0-9A-Za-z!#$%&+\./:;?@^_|~-]|[\n\t]/g;
  93. nSeqName = invReplace(regex_sam_restriction, sequenceName);
  94. return [4 /*yield*/, fs_1.default.promises.writeFile(filePath, '>' + nSeqName + '\n' + ((_a = sequence.match(r)) === null || _a === void 0 ? void 0 : _a.join('\n').toUpperCase()))];
  95. case 1:
  96. _b.sent();
  97. resolve(true);
  98. return [3 /*break*/, 3];
  99. case 2:
  100. error_1 = _b.sent();
  101. console.log(error_1);
  102. reject(false);
  103. return [3 /*break*/, 3];
  104. case 3: return [2 /*return*/];
  105. }
  106. });
  107. }); })];
  108. });
  109. });
  110. };
  111. exports.writeSequence = writeSequence;
  112. var makeReference = function (sequenceName, sequence, filePath, lineN) {
  113. if (lineN === void 0) { lineN = 80; }
  114. return __awaiter(void 0, void 0, void 0, function () {
  115. return __generator(this, function (_a) {
  116. switch (_a.label) {
  117. case 0: return [4 /*yield*/, writeSequence(sequenceName, sequence, filePath, lineN)];
  118. case 1:
  119. if (!_a.sent()) return [3 /*break*/, 3];
  120. return [4 /*yield*/, async_exec('bwa', ['index', filePath], function () { return console.log; })];
  121. case 2:
  122. _a.sent();
  123. _a.label = 3;
  124. case 3: return [2 /*return*/];
  125. }
  126. });
  127. });
  128. };
  129. exports.makeReference = makeReference;
  130. var asyncBwaMem = function (refPath, reads,
  131. // R1 : string | Array<string>,
  132. // R2 : string | Array<string>,
  133. runName, libName, outputDir, onData, options) {
  134. return new Promise(function (resolve, reject) { return __awaiter(void 0, void 0, void 0, function () {
  135. var defaultOptions, refName, bwa, samblaster, samtools, sambamba, readsIn, isPairedEnd, R1, R2, R1_arr, R2_arr, R1_kitty, R2_kitty, R1_in, R2_in, bam, bamSorted, retObj, threads, samblasterCmd, discordantFile, splitterFile, unmappedFile, code, code_sort, err_1;
  136. return __generator(this, function (_a) {
  137. switch (_a.label) {
  138. case 0:
  139. _a.trys.push([0, 6, , 7]);
  140. defaultOptions = {
  141. output_discordant: true,
  142. output_splitted: true,
  143. output_unmapped: true
  144. };
  145. if (typeof options === 'undefined') {
  146. options = defaultOptions;
  147. }
  148. else {
  149. options = __assign(__assign({}, defaultOptions), options);
  150. }
  151. refName = path_1.default.parse(refPath).name;
  152. bwa = 'bwa';
  153. samblaster = 'samblaster';
  154. samtools = 'samtools';
  155. sambamba = 'sambamba';
  156. readsIn = void 0;
  157. isPairedEnd = false;
  158. if (Array.isArray(reads)) {
  159. isPairedEnd = true;
  160. console.log('Assuming paired end reads');
  161. R1 = reads[0], R2 = reads[1];
  162. R1_arr = Array.isArray(R1) ? R1.join(' ') : R1;
  163. R2_arr = Array.isArray(R2) ? R2.join(' ') : R2;
  164. R1_kitty = R1_arr.slice(-2) === 'gz' ? 'zcat' : 'cat';
  165. R2_kitty = R2_arr.slice(-2) === 'gz' ? 'zcat' : 'cat';
  166. R1_in = "'< ".concat(R1_kitty, " ").concat(R1_arr, "'");
  167. R2_in = "'< ".concat(R2_kitty, " ").concat(R2_arr, "'");
  168. readsIn = R1_in + ' ' + R2_in;
  169. }
  170. else {
  171. readsIn = reads;
  172. }
  173. bam = path_1.default.join(outputDir, "bwa_mem_properly_on_".concat(refName, ".bam"));
  174. bamSorted = path_1.default.join(outputDir, "bwa_mem_properly_on_".concat(refName, ".sorted.bam"));
  175. retObj = { bamSorted: bamSorted };
  176. if (options === null || options === void 0 ? void 0 : options.remove_mapped) {
  177. bam = '/dev/null';
  178. delete retObj.bamSorted;
  179. }
  180. threads = String((0, os_1.cpus)().length);
  181. samblasterCmd = [];
  182. // https://github.com/GregoryFaust/samblaster
  183. samblasterCmd = ['|', samblaster,
  184. '--addMateTags',
  185. '-a',
  186. '-e', // Exclude reads marked as duplicates from discordant, splitter, and/or unmapped
  187. ];
  188. if ((options === null || options === void 0 ? void 0 : options.output_discordant) || (options === null || options === void 0 ? void 0 : options.output_splitted)) {
  189. console.log('Using samblaster');
  190. if (options === null || options === void 0 ? void 0 : options.output_discordant) {
  191. if (!isPairedEnd) {
  192. console.log('Discordant reads can be found only in paired reads, skipping');
  193. }
  194. else {
  195. discordantFile = path_1.default.join(outputDir, "bwa_mem_discordants_on_".concat(refName, ".sam"));
  196. console.log('Discordant reads file path: ', discordantFile);
  197. samblasterCmd = __spreadArray(__spreadArray([], samblasterCmd, true), ['-d', discordantFile], false);
  198. retObj = __assign(__assign({}, retObj), { discordantFile: discordantFile });
  199. }
  200. }
  201. if (!isPairedEnd) {
  202. samblasterCmd = __spreadArray(__spreadArray([], samblasterCmd, true), ['--ignoreUnmated'], false);
  203. }
  204. if (options === null || options === void 0 ? void 0 : options.output_splitted) {
  205. splitterFile = path_1.default.join(outputDir, "bwa_mem_splitters_on_".concat(refName, ".sam"));
  206. console.log('Splitted reads file path: ', splitterFile);
  207. samblasterCmd = __spreadArray(__spreadArray([], samblasterCmd, true), ['-s', splitterFile], false);
  208. retObj = __assign(__assign({}, retObj), { splitterFile: splitterFile });
  209. }
  210. }
  211. if (options === null || options === void 0 ? void 0 : options.output_unmapped) {
  212. unmappedFile = path_1.default.join(outputDir, "bwa_mem_unmapped_on_".concat(refName, ".fq"));
  213. console.log('Unmapped reads file path: ', unmappedFile);
  214. samblasterCmd = __spreadArray(__spreadArray([], samblasterCmd, true), ['-u', unmappedFile], false);
  215. retObj = __assign(__assign({}, retObj), { unmappedFile: unmappedFile });
  216. }
  217. if (!!fs_1.default.existsSync(refPath + '.amb')) return [3 /*break*/, 2];
  218. return [4 /*yield*/, async_exec(bwa, ['index', refPath], function (message) { return onData('[BWA-INDEX] ' + message); })];
  219. case 1:
  220. _a.sent();
  221. _a.label = 2;
  222. case 2:
  223. console.log(options, samblasterCmd);
  224. return [4 /*yield*/, async_exec(bwa, __spreadArray(__spreadArray(['mem',
  225. '-t', threads,
  226. '-R', "\"@RG\\tPL:Illumina\\tID:".concat(+(new Date), "\\tSM:").concat(runName, "\\tLB:").concat(libName, "\""), refPath,
  227. readsIn], samblasterCmd, true), ['|',
  228. samtools,
  229. 'view',
  230. '-Sb',
  231. '-',
  232. '>',
  233. bam], false), function (message) { return onData('[BWA-MEM] ' + message); })];
  234. case 3:
  235. code = _a.sent();
  236. onData('[BWA-MEM][EXIT CODE] ' + code);
  237. if (!retObj.bamSorted) return [3 /*break*/, 5];
  238. return [4 /*yield*/, async_exec(sambamba, ['sort',
  239. '-t', threads,
  240. bam
  241. ], function (message) { return onData('[SAMBAMBA-SORT] ' + message); })];
  242. case 4:
  243. code_sort = _a.sent();
  244. onData('[SAMBAMBA-SORT][EXIT CODE] ' + code_sort);
  245. fs_1.default.unlinkSync(bam);
  246. _a.label = 5;
  247. case 5:
  248. resolve(retObj);
  249. return [3 /*break*/, 7];
  250. case 6:
  251. err_1 = _a.sent();
  252. reject(err_1);
  253. return [3 /*break*/, 7];
  254. case 7: return [2 /*return*/];
  255. }
  256. });
  257. }); });
  258. };
  259. exports.asyncBwaMem = asyncBwaMem;