Mercurial > repos > cstrittmatter > mitokmer
comparison kmer_read_m3.cpp @ 0:1472b4f4fbfe draft
planemo upload commit a4aeec0f52695b1cbda85ed1faf85d66ba897ab2
| author | cstrittmatter |
|---|---|
| date | Fri, 12 Apr 2019 14:58:18 -0400 |
| parents | |
| children |
comparison
equal
deleted
inserted
replaced
| -1:000000000000 | 0:1472b4f4fbfe |
|---|---|
| 1 // newkmer_5.cpp load k-mers for all species | |
| 2 // test reads by SmithWaterman | |
| 3 // MKM 1 APR 2016 | |
| 4 //-std=c++0x -D__NO_INLINE__ ... at end: -lz | |
| 5 //#include "stdafx.h" | |
| 6 #include <fstream> | |
| 7 #include <stdlib.h> | |
| 8 #include <stdio.h> | |
| 9 #include <math.h> | |
| 10 #include <iostream> | |
| 11 #include <sstream> | |
| 12 #include <iomanip> | |
| 13 #include <ostream> | |
| 14 #include <string> | |
| 15 #include <vector> | |
| 16 #include <unordered_set> | |
| 17 #include <time.h> | |
| 18 //#include <sys/time.h> | |
| 19 #include <cstddef> | |
| 20 #include <cstring> | |
| 21 #include <algorithm> | |
| 22 #include <set> | |
| 23 #include <zlib.h> | |
| 24 #include <dirent.h> | |
| 25 using namespace std; | |
| 26 | |
| 27 const int minalign = 0; //120; | |
| 28 unordered_set<int> myset = { 0}; | |
| 29 | |
| 30 typedef int ptype; | |
| 31 typedef unsigned long long ui64; | |
| 32 typedef unsigned long long ktype; | |
| 33 typedef unsigned int vtype; | |
| 34 typedef unsigned short int otype; | |
| 35 typedef unsigned long long itype; | |
| 36 | |
| 37 const int KSIZE = 30; //kmer size | |
| 38 const int MAXREP = 2048; | |
| 39 const int REPSHIFT = 11; //(log of maxrep) | |
| 40 const int SAVENUM = 12; //save this many reads | |
| 41 const itype MAXHASH = (1 << 30); //pow2 size of hashtable | |
| 42 const int MAXREPROBE = 16; | |
| 43 const int GAPO = 11; | |
| 44 const int GAPX = 1; | |
| 45 const int MATCH = 5; | |
| 46 const int MISMATCH = -4; | |
| 47 const bool INIGAPPEN = false; | |
| 48 const int MAXALIGN = 1024; | |
| 49 const int beam = 8; | |
| 50 | |
| 51 vector<string> accession; | |
| 52 vector<int> targno; | |
| 53 int mcount, savepos, tct, num_orgs=0, save_target=0; | |
| 54 vector<int> gcount, ucount; | |
| 55 ofstream outread; | |
| 56 set<ktype> kmer_seen; | |
| 57 string fdir = ""; //directory for fasta.gz files | |
| 58 | |
| 59 const char bases[4] = { 'A', 'C', 'G', 'T' }; | |
| 60 const ui64 one = 1; | |
| 61 const ui64 two = 2; | |
| 62 const ui64 three = 3; | |
| 63 const ui64 mask = (one << (KSIZE * 2)) - 1; | |
| 64 const ui64 hiC = one << (KSIZE - 1) * 2; | |
| 65 const ui64 hiG = two << (KSIZE - 1) * 2; | |
| 66 const ui64 hiT = three << (KSIZE - 1) * 2; | |
| 67 const ui64 hi1 = one << 62; | |
| 68 const ui64 hi2 = two << 62; | |
| 69 const ui64 hi3 = three << 62; | |
| 70 const ui64 himask = hi1 - one; | |
| 71 | |
| 72 const unsigned BUFLEN = 0x4000; | |
| 73 | |
| 74 void error(const char* const msg) | |
| 75 { | |
| 76 cerr << msg << "\n"; | |
| 77 exit(255); | |
| 78 } | |
| 79 | |
| 80 string int2str(ktype key) | |
| 81 { | |
| 82 string sequence = ""; | |
| 83 for (int i = 0; i < KSIZE; ++i) | |
| 84 { | |
| 85 sequence = bases[key & 3] + sequence; | |
| 86 key >>= 2; | |
| 87 } | |
| 88 return sequence; | |
| 89 } | |
| 90 | |
| 91 class Tree1 | |
| 92 { | |
| 93 public: | |
| 94 | |
| 95 vector<vector<int>> children; | |
| 96 vector<int> parent; | |
| 97 int root; | |
| 98 | |
| 99 Tree1(int nt) | |
| 100 { | |
| 101 vector<int> blank; | |
| 102 root = 1; | |
| 103 for (int i = 0; i < nt; i++) | |
| 104 { | |
| 105 parent.push_back(root); | |
| 106 children.push_back(blank); | |
| 107 } | |
| 108 } | |
| 109 | |
| 110 ~Tree1() | |
| 111 { | |
| 112 } | |
| 113 | |
| 114 void add_edge(int x, int y) | |
| 115 { | |
| 116 parent[y] = x; | |
| 117 children[x].push_back(y); | |
| 118 } | |
| 119 | |
| 120 int msca(int x, int y) | |
| 121 { | |
| 122 //return most specific common ancestor of nodes x and y | |
| 123 set<int> ancestors; | |
| 124 int z; | |
| 125 | |
| 126 ancestors.insert(root); | |
| 127 z = x; | |
| 128 //cout << x << " " << y << " " << z << endl; | |
| 129 while (z != root) | |
| 130 { | |
| 131 ancestors.insert(z); | |
| 132 z = get_parent(z); | |
| 133 //cout << x << " " << y << " " << z << endl; | |
| 134 } | |
| 135 z = y; | |
| 136 if (ancestors.find(y) != ancestors.end()) | |
| 137 return x; //x is descendent of y | |
| 138 while (ancestors.find(z) == ancestors.end()) | |
| 139 { | |
| 140 z = get_parent(z); | |
| 141 if (z==x) //y is descendent of x | |
| 142 return y; | |
| 143 //cout << x << " " << y << " " << z << endl; | |
| 144 } | |
| 145 return z; //common ancesctor | |
| 146 } | |
| 147 | |
| 148 int get_parent(int x) | |
| 149 { | |
| 150 if (x != root && x > 0) | |
| 151 return parent[x]; | |
| 152 else | |
| 153 return root; | |
| 154 } | |
| 155 | |
| 156 }; | |
| 157 | |
| 158 Tree1 *taxonomy; | |
| 159 | |
| 160 class Hashtable | |
| 161 { | |
| 162 public: | |
| 163 | |
| 164 itype size; | |
| 165 | |
| 166 struct Cell | |
| 167 { | |
| 168 ktype key; | |
| 169 vtype value; | |
| 170 int org; | |
| 171 ptype position; | |
| 172 bool fstrand; | |
| 173 } *pHash; | |
| 174 | |
| 175 Hashtable() | |
| 176 { | |
| 177 size = 0; | |
| 178 pHash = new Cell[MAXHASH]; | |
| 179 HashClear(); | |
| 180 //for (int i = 1; i<MAXREPROBE; i++) | |
| 181 // reprobeVal[i] = reprobeVal[i - 1] + i; //i*i/2+i/2 | |
| 182 } | |
| 183 | |
| 184 ~Hashtable() | |
| 185 { | |
| 186 delete[] pHash; | |
| 187 size = 0; | |
| 188 } | |
| 189 | |
| 190 // from code.google.com/p/smhasher/wiki/MurmurHash3 | |
| 191 itype integerHash(ui64 k) | |
| 192 { | |
| 193 k ^= k >> 33; | |
| 194 k *= 0xff51afd7ed558ccd; | |
| 195 k ^= k >> 33; | |
| 196 k *= 0xc4ceb9fe1a85ec53; | |
| 197 k ^= k >> 33; | |
| 198 return (itype) k; | |
| 199 } | |
| 200 | |
| 201 void HashClear() | |
| 202 { | |
| 203 memset(pHash, 0, MAXHASH * sizeof(Cell)); | |
| 204 } | |
| 205 | |
| 206 int getHash(ktype key, otype &org, ptype &position, bool &fstrand) | |
| 207 { | |
| 208 //0 target=bad | |
| 209 //returns target if probe present in table | |
| 210 //sets position and fstrand | |
| 211 itype index, reprobe, i, hash; | |
| 212 | |
| 213 position = 0; | |
| 214 hash = integerHash(key); | |
| 215 reprobe = 0; | |
| 216 i = 0; | |
| 217 do | |
| 218 { | |
| 219 index = ((hash + reprobe) & (MAXHASH-1)); //for power of 2 size | |
| 220 reprobe += ++i; | |
| 221 if (pHash[index].value == 0) | |
| 222 { //empty cell | |
| 223 return 0; | |
| 224 } | |
| 225 else if (pHash[index].key == key) | |
| 226 { | |
| 227 position = pHash[index].position; | |
| 228 fstrand = pHash[index].fstrand; | |
| 229 org = pHash[index].org; | |
| 230 return (pHash[index].value); | |
| 231 } | |
| 232 } while (reprobe < MAXHASH && i < MAXREPROBE); | |
| 233 | |
| 234 return 0; | |
| 235 } | |
| 236 | |
| 237 void add_kmer(ktype key, vtype target, otype org, ptype position, bool fstrand) | |
| 238 { //add kmer to table | |
| 239 itype index, reprobe, i, hash; | |
| 240 bool notdone=true; | |
| 241 | |
| 242 hash = integerHash(key); | |
| 243 reprobe = 0; | |
| 244 i = 0; | |
| 245 do | |
| 246 { | |
| 247 index = ((hash + reprobe) & (MAXHASH-1)); //for power of 2 size | |
| 248 reprobe += ++i; | |
| 249 | |
| 250 if (pHash[index].value == 0) | |
| 251 { //empty cell | |
| 252 notdone = false; | |
| 253 pHash[index].key = key; | |
| 254 pHash[index].value = target; | |
| 255 pHash[index].org = org; | |
| 256 pHash[index].position = position; | |
| 257 pHash[index].fstrand = fstrand; | |
| 258 if (++size > MAXHASH - 32) | |
| 259 { | |
| 260 cout << "out of memory in table " << endl; | |
| 261 exit(1); | |
| 262 } | |
| 263 } | |
| 264 } while (notdone); | |
| 265 } | |
| 266 | |
| 267 }; //class Hashtable | |
| 268 Hashtable *ht; | |
| 269 | |
| 270 string process_seq(string seq) | |
| 271 { | |
| 272 //format nucleotide sequence to all caps ACTG, all else is N | |
| 273 string seq2; | |
| 274 string::iterator it1; | |
| 275 char base; | |
| 276 | |
| 277 seq2 = ""; | |
| 278 for (it1 = seq.begin(); it1 != seq.end(); ++it1) | |
| 279 { | |
| 280 base = *it1; | |
| 281 switch (base) | |
| 282 { | |
| 283 case 'a': | |
| 284 seq2 += "A"; | |
| 285 break; | |
| 286 case 'c': | |
| 287 seq2 += "C"; | |
| 288 break; | |
| 289 case 'g': | |
| 290 seq2 += "G"; | |
| 291 break; | |
| 292 case 't': | |
| 293 seq2 += "T"; | |
| 294 break; | |
| 295 case 'A': | |
| 296 seq2 += "A"; | |
| 297 break; | |
| 298 case 'C': | |
| 299 seq2 += "C"; | |
| 300 break; | |
| 301 case 'G': | |
| 302 seq2 += "G"; | |
| 303 break; | |
| 304 case 'T': | |
| 305 seq2 += "T"; | |
| 306 break; | |
| 307 default: | |
| 308 seq2 += "N"; | |
| 309 break; | |
| 310 } | |
| 311 } | |
| 312 | |
| 313 return seq2; | |
| 314 } | |
| 315 | |
| 316 string load_data3(gzFile in) | |
| 317 { //read gzip fasta file and extract nt sequence | |
| 318 char buf[BUFLEN]; | |
| 319 char* offset = buf; | |
| 320 string line; | |
| 321 string sequence = ""; | |
| 322 | |
| 323 while (true) | |
| 324 { | |
| 325 int err, len = sizeof(buf)-(offset-buf); | |
| 326 if (len == 0) error("Buffer to small for input line lengths"); | |
| 327 | |
| 328 len = gzread(in, offset, len); | |
| 329 | |
| 330 if (len == 0) break; | |
| 331 if (len < 0) error(gzerror(in, &err)); | |
| 332 | |
| 333 char* cur = buf; | |
| 334 char* end = offset+len; | |
| 335 | |
| 336 for (char* eol; (cur<end) && (eol = find(cur, end, '\n')) < end; cur = eol + 1) | |
| 337 { | |
| 338 line = string(cur, eol); | |
| 339 if (line.back() == '\r') | |
| 340 line.pop_back(); | |
| 341 if (line.length() > 0) | |
| 342 { | |
| 343 if (line.at(0) == '>') | |
| 344 { | |
| 345 sequence += "N"; //contig separator | |
| 346 } | |
| 347 else | |
| 348 { | |
| 349 sequence += process_seq(line); | |
| 350 } | |
| 351 } | |
| 352 //cout << line << endl; | |
| 353 } | |
| 354 | |
| 355 // any trailing data in [eol, end) now is a partial line | |
| 356 offset = copy(cur, end, buf); | |
| 357 } | |
| 358 | |
| 359 // trailing data without eol | |
| 360 line = string(buf, offset); | |
| 361 if (gzclose(in) != Z_OK) error("failed gzclose"); | |
| 362 | |
| 363 return sequence; | |
| 364 } | |
| 365 | |
| 366 int *tableM, *tableI, *tableD; | |
| 367 | |
| 368 void tableinit() | |
| 369 { | |
| 370 int i,j; | |
| 371 const int NINF = -2000000000; | |
| 372 | |
| 373 tableM[0] = 0; | |
| 374 tableI[0] = NINF; | |
| 375 tableD[0] = NINF; | |
| 376 for (i = 1; i<MAXALIGN; i++) //initialize upper row | |
| 377 { | |
| 378 j = (0 > i - beam - 1) ? 0 : i - beam - 1; | |
| 379 tableI[j*MAXALIGN + i] = NINF; | |
| 380 if (INIGAPPEN) | |
| 381 { | |
| 382 tableD[j*MAXALIGN + i] = -GAPO - (i - 1)*GAPX; | |
| 383 tableM[j*MAXALIGN + i] = -GAPO - (i - 1)*GAPX; | |
| 384 } | |
| 385 else | |
| 386 { | |
| 387 tableD[0 + i] = 0; | |
| 388 tableM[0 + i] = 0; | |
| 389 } | |
| 390 } | |
| 391 for (j = 1; j<MAXALIGN; j++) //init left column | |
| 392 { | |
| 393 i = (0 > j - beam - 1) ? 0 : j - beam - 1; | |
| 394 tableD[j*MAXALIGN + i] = NINF; | |
| 395 if (INIGAPPEN) | |
| 396 { | |
| 397 tableI[j*MAXALIGN + i] = -GAPO - (j - 1)*GAPX; | |
| 398 tableM[j*MAXALIGN + i] = -GAPO - (j - 1)*GAPX; | |
| 399 } | |
| 400 else | |
| 401 { | |
| 402 tableI[j*MAXALIGN + i] = 0; | |
| 403 tableM[j*MAXALIGN + i] = 0; | |
| 404 } | |
| 405 } | |
| 406 } | |
| 407 | |
| 408 int align(string dna1, string dna2) | |
| 409 { | |
| 410 //smith waterman alignment with beam | |
| 411 //returns alignment score | |
| 412 int b1, b2, len1, len2, tindex; | |
| 413 int b,bb,c,cc; | |
| 414 int i, j, maxval, startx, starty; | |
| 415 int i1, i2; | |
| 416 | |
| 417 len1 = dna1.length() + 1; | |
| 418 len2 = dna2.length() + 1; | |
| 419 | |
| 420 j=1; | |
| 421 do | |
| 422 { | |
| 423 i = (j <= beam) ? j : j - beam; | |
| 424 i2 = (j + beam > len1) ? len1 : j + beam; | |
| 425 do | |
| 426 { | |
| 427 maxval = max(max(tableM[(j-1)*MAXALIGN + i-1], tableI[(j-1)*MAXALIGN + i-1]), tableD[(j-1)*MAXALIGN + i-1]); | |
| 428 //copy/replace | |
| 429 if (dna1[i-1] != dna2[j-1]) | |
| 430 maxval += MISMATCH; | |
| 431 else maxval += MATCH; | |
| 432 tableM[j*MAXALIGN + i] = maxval; | |
| 433 | |
| 434 b=tableM[(j-1)*MAXALIGN + i] - GAPO; //new gap | |
| 435 bb = tableI[(j-1)*MAXALIGN + i] -GAPX; //extend | |
| 436 tableI[j*MAXALIGN + i] = (bb<b) ? b : bb; | |
| 437 | |
| 438 c=tableM[j*MAXALIGN + i-1] - GAPO; //new gap | |
| 439 cc = tableD[j*MAXALIGN + i-1] -GAPX; //extend | |
| 440 tableD[j*MAXALIGN + i] = (cc<c) ? c : cc; | |
| 441 } while(++i < i2); | |
| 442 } while(++j < len2); | |
| 443 | |
| 444 i = len1 - 1; | |
| 445 j = len2 - 1; //process last entry | |
| 446 | |
| 447 maxval = max(max(tableM[j*MAXALIGN + i], tableI[j*MAXALIGN + i]), tableD[j*MAXALIGN + i]); | |
| 448 | |
| 449 return maxval; | |
| 450 | |
| 451 } | |
| 452 | |
| 453 int process_read(string &sequence, string acc, int start, int stop) | |
| 454 { | |
| 455 //returns id of read, 0 = no match, updates gcount, ucount | |
| 456 //will do align for first minalign of each target | |
| 457 //will save first SAVENUM reads of each target | |
| 458 | |
| 459 char base; | |
| 460 int it1, it2; | |
| 461 string kmer, sequence2, dna1, dna1r, dna2, fname, finalkeystr; | |
| 462 ktype keyF, keyR, key; | |
| 463 int i, cpos, minscr, gpos; | |
| 464 bool fstrand1, fstrand2; | |
| 465 int readlength = stop - start + 1, readlen2; | |
| 466 int st2, scr, stlen2; | |
| 467 ktype target, final_targ=0; | |
| 468 ptype position; | |
| 469 otype org; | |
| 470 int maxct, totct, rtarg=0; | |
| 471 bool reject = false; | |
| 472 | |
| 473 cpos = 0; | |
| 474 keyF = 0; | |
| 475 keyR = 0; | |
| 476 gpos = 0; | |
| 477 minscr = 5 * sequence.length() / 2; | |
| 478 for (it1 = start; it1 <= stop; ++it1) | |
| 479 { | |
| 480 base = sequence.at(it1); | |
| 481 switch (base) | |
| 482 { | |
| 483 case 'A': | |
| 484 keyF = (keyF << 2) & mask; | |
| 485 keyR = (keyR >> 2) | hiT; | |
| 486 cpos++; | |
| 487 break; | |
| 488 case 'C': | |
| 489 keyF = ((keyF << 2) & mask) | 1; | |
| 490 keyR = (keyR >> 2) | hiG; | |
| 491 cpos++; | |
| 492 break; | |
| 493 case 'G': | |
| 494 keyF = ((keyF << 2) & mask) | 2; | |
| 495 keyR = (keyR >> 2) | hiC; | |
| 496 cpos++; | |
| 497 break; | |
| 498 case 'T': | |
| 499 keyF = ((keyF << 2) & mask) | 3; | |
| 500 keyR = keyR >> 2; | |
| 501 cpos++; | |
| 502 break; | |
| 503 case 'a': | |
| 504 keyF = (keyF << 2) & mask; | |
| 505 keyR = (keyR >> 2) | hiT; | |
| 506 cpos++; | |
| 507 break; | |
| 508 case 'c': | |
| 509 keyF = ((keyF << 2) & mask) | 1; | |
| 510 keyR = (keyR >> 2) | hiG; | |
| 511 cpos++; | |
| 512 break; | |
| 513 case 'g': | |
| 514 keyF = ((keyF << 2) & mask) | 2; | |
| 515 keyR = (keyR >> 2) | hiC; | |
| 516 cpos++; | |
| 517 break; | |
| 518 case 't': | |
| 519 keyF = ((keyF << 2) & mask) | 3; | |
| 520 keyR = keyR >> 2; | |
| 521 cpos++; | |
| 522 break; | |
| 523 default: | |
| 524 cpos = 0; //ambiguous character | |
| 525 keyF = 0; | |
| 526 keyR = 0; | |
| 527 break; | |
| 528 } | |
| 529 gpos++; | |
| 530 if (cpos == KSIZE) | |
| 531 { | |
| 532 key = keyF < keyR ? keyF : keyR; | |
| 533 target = ht->getHash(key, org, position, fstrand2); | |
| 534 //position is ending base, 1-based | |
| 535 if (target > 0 && gcount[target] < minalign && target != final_targ) | |
| 536 { //do alignment on initial reads of a species | |
| 537 fname = fdir + accession[org] + ".fasta.gz"; | |
| 538 sequence2 = load_data3(gzopen(fname.c_str(), "rb")); | |
| 539 stlen2 = sequence2.length(); | |
| 540 fstrand1 = (keyF < keyR); | |
| 541 readlen2 = readlength; | |
| 542 if (fstrand1 == fstrand2) | |
| 543 { | |
| 544 st2 = position - it1; | |
| 545 if (st2 < 0) st2 = 0; | |
| 546 if (st2 + readlen2 > stlen2) readlen2 = stlen2 - st2; | |
| 547 dna1 = sequence.substr(start, readlength); | |
| 548 dna2 = sequence2.substr(st2, readlen2); | |
| 549 } | |
| 550 else | |
| 551 { | |
| 552 //reverse complement dna1 | |
| 553 st2 = position - KSIZE + 2 + it1 - readlength; | |
| 554 if (st2 < 0) st2 = 0; | |
| 555 if (st2 + readlen2 > stlen2) readlen2 = stlen2 - st2; | |
| 556 //dna1 = sequence.substr(start, readlength); | |
| 557 dna1 = ""; | |
| 558 for (it2 = stop; it2 >= start; --it2) | |
| 559 { | |
| 560 base = sequence.at(it2); | |
| 561 switch (base) | |
| 562 { | |
| 563 case 'A': | |
| 564 dna1.push_back('T'); | |
| 565 break; | |
| 566 case 'C': | |
| 567 dna1.push_back('G'); | |
| 568 break; | |
| 569 case 'G': | |
| 570 dna1.push_back('C'); | |
| 571 break; | |
| 572 case 'T': | |
| 573 dna1.push_back('A'); | |
| 574 break; | |
| 575 default: | |
| 576 dna1.push_back('N'); | |
| 577 break; | |
| 578 } | |
| 579 } | |
| 580 dna2 = sequence2.substr(st2, readlen2); | |
| 581 } | |
| 582 scr = align(dna1, dna2); | |
| 583 if (scr < minscr) | |
| 584 { | |
| 585 rtarg = target; | |
| 586 target = 0; //fail | |
| 587 reject = true; | |
| 588 //cout << rtarg << ": failed " << scr << endl; | |
| 589 } | |
| 590 } | |
| 591 if (final_targ > 0 && target > 0) | |
| 592 { | |
| 593 final_targ = taxonomy->msca(target, final_targ); | |
| 594 finalkeystr = int2str(key); | |
| 595 } | |
| 596 else if (target > 0) | |
| 597 { | |
| 598 final_targ = target; | |
| 599 finalkeystr = int2str(key); | |
| 600 } | |
| 601 if (target > 1) | |
| 602 { | |
| 603 if (kmer_seen.find(key) == kmer_seen.end()) | |
| 604 { | |
| 605 ucount[target]++; | |
| 606 kmer_seen.insert(key); | |
| 607 } | |
| 608 } | |
| 609 cpos--; | |
| 610 } | |
| 611 } //it | |
| 612 /* | |
| 613 if (final_targ > 1 && gcount[final_targ] < SAVENUM && save_target == 0) | |
| 614 { | |
| 615 outread << ">" << final_targ << ":" << acc << endl << sequence.substr(start, stop-start+1) << endl; | |
| 616 outread << "[" << finalkeystr << "]" << endl; | |
| 617 } | |
| 618 if (final_targ > 1 && final_targ == save_target) | |
| 619 { | |
| 620 outread << ">" << final_targ << ":" << acc << endl << sequence.substr(start, stop-start+1) << endl; | |
| 621 } */ | |
| 622 gcount[final_targ]++; | |
| 623 tct++; | |
| 624 | |
| 625 return final_targ; | |
| 626 } | |
| 627 | |
| 628 | |
| 629 void process_kmer(string sequence, vtype target, otype org, ptype position, bool fstrand) | |
| 630 { //adds kmer from datafile into hashtable | |
| 631 char base; | |
| 632 string::iterator it1; | |
| 633 int cpos, gpos; | |
| 634 ui64 keyF; | |
| 635 | |
| 636 cpos = 0; | |
| 637 gpos = 0; | |
| 638 keyF = 0; | |
| 639 | |
| 640 for (it1 = sequence.begin(); it1 != sequence.end(); ++it1) | |
| 641 { | |
| 642 base = *it1; | |
| 643 switch (base) | |
| 644 { | |
| 645 case 'A': | |
| 646 keyF = (keyF << 2) & mask; | |
| 647 cpos++; | |
| 648 break; | |
| 649 case 'C': | |
| 650 keyF = ((keyF << 2) & mask) | 1; | |
| 651 cpos++; | |
| 652 break; | |
| 653 case 'G': | |
| 654 keyF = ((keyF << 2) & mask) | 2; | |
| 655 cpos++; | |
| 656 break; | |
| 657 case 'T': | |
| 658 keyF = ((keyF << 2) & mask) | 3; | |
| 659 cpos++; | |
| 660 break; | |
| 661 default: | |
| 662 cpos = 0; //ambiguous character | |
| 663 keyF = 0; | |
| 664 break; | |
| 665 } | |
| 666 if (cpos == KSIZE) | |
| 667 { | |
| 668 ht->add_kmer(keyF, target, org, position, fstrand); | |
| 669 cpos--; | |
| 670 } | |
| 671 gpos++; | |
| 672 } //it | |
| 673 | |
| 674 } | |
| 675 | |
| 676 void process_qual(string acc, string &seq, string &qual) | |
| 677 { //uses PHRED 33 scores | |
| 678 //to trim read and pass it to process_read | |
| 679 const int cutoff_qual = 17; | |
| 680 const char cutoff_char = 32 + cutoff_qual; | |
| 681 int start, stop, target, i; | |
| 682 int window_size = 4; | |
| 683 int window_val, window_cut = cutoff_qual * window_size; | |
| 684 string trim_seq; | |
| 685 | |
| 686 stop = seq.length()-1; | |
| 687 start = 0; | |
| 688 //trim all low qual bases from end | |
| 689 while (qual.at(start) < cutoff_char && start < stop) | |
| 690 start++; | |
| 691 while (qual.at(stop) < cutoff_char && stop > start) | |
| 692 stop--; | |
| 693 //trim by window | |
| 694 if (start < stop - window_size) | |
| 695 { | |
| 696 window_val = 0; | |
| 697 for (i=0; i<window_size; i++) | |
| 698 window_val += qual.at(start+i) - 32; | |
| 699 while (window_val < window_cut && start < stop - window_size) | |
| 700 { | |
| 701 window_val += qual.at(start+window_size) - qual.at(start); | |
| 702 start++; | |
| 703 } | |
| 704 } | |
| 705 if (start < stop - window_size) | |
| 706 { | |
| 707 window_val = 0; | |
| 708 for (i=0; i<window_size; i++) | |
| 709 window_val += qual.at(stop-i) - 32; | |
| 710 while (window_val < window_cut && start < stop - window_size) | |
| 711 { | |
| 712 window_val += qual.at(stop-window_size) - qual.at(stop); | |
| 713 stop--; | |
| 714 } | |
| 715 } | |
| 716 //trim_seq = seq.substr(start,stop-start+1); | |
| 717 if (stop - start >= KSIZE) | |
| 718 { | |
| 719 target = process_read(seq,acc,start,stop); | |
| 720 } | |
| 721 | |
| 722 } | |
| 723 | |
| 724 void process_fqgz(gzFile in) | |
| 725 { //read gzip fastq file and extract sequence and quality lines | |
| 726 //passes them to process_qual | |
| 727 char buf[BUFLEN]; | |
| 728 char* offset = buf; | |
| 729 string line, seq, acc; | |
| 730 int mod4=0; | |
| 731 | |
| 732 while (true) | |
| 733 { | |
| 734 int err, len = sizeof(buf)-(offset-buf); | |
| 735 if (len == 0) error("Buffer to small for input line lengths"); | |
| 736 | |
| 737 len = gzread(in, offset, len); | |
| 738 | |
| 739 if (len == 0) break; | |
| 740 if (len < 0) error(gzerror(in, &err)); | |
| 741 | |
| 742 char* cur = buf; | |
| 743 char* end = offset+len; | |
| 744 | |
| 745 for (char* eol; (cur<end) && (eol = find(cur, end, '\n')) < end; cur = eol + 1) | |
| 746 { | |
| 747 line = string(cur, eol); | |
| 748 if (line.back() == '\r') //windows | |
| 749 line.pop_back(); | |
| 750 if (line.length() > 0) | |
| 751 { | |
| 752 if (mod4 == 1) | |
| 753 { | |
| 754 seq = line; | |
| 755 } | |
| 756 else if (mod4 == 0) | |
| 757 { | |
| 758 acc = line; | |
| 759 } | |
| 760 else if (mod4 == 3) | |
| 761 { | |
| 762 process_qual(acc, seq, line); | |
| 763 } | |
| 764 mod4 = (mod4 + 1) % 4; | |
| 765 | |
| 766 } | |
| 767 //cout << line << endl; | |
| 768 } | |
| 769 | |
| 770 // any trailing data in [eol, end) now is a partial line | |
| 771 offset = copy(cur, end, buf); | |
| 772 } | |
| 773 | |
| 774 // trailing data without eol | |
| 775 line = string(buf, offset); | |
| 776 | |
| 777 if (gzclose(in) != Z_OK) error("failed gzclose"); | |
| 778 } | |
| 779 | |
| 780 void process_fagz(gzFile in) | |
| 781 { //read gzip fasta file and extract sequence | |
| 782 //passes read to process_read | |
| 783 char buf[BUFLEN]; | |
| 784 char* offset = buf; | |
| 785 string line; | |
| 786 string sequence = "", acc = ""; | |
| 787 int c1, c2, target; | |
| 788 | |
| 789 cout << "true" << endl; | |
| 790 | |
| 791 while (true) | |
| 792 { | |
| 793 int err, len = sizeof(buf)-(offset-buf); | |
| 794 if (len == 0) error("Buffer to small for input line lengths"); | |
| 795 | |
| 796 len = gzread(in, offset, len); | |
| 797 | |
| 798 if (len == 0) break; | |
| 799 if (len < 0) error(gzerror(in, &err)); | |
| 800 | |
| 801 char* cur = buf; | |
| 802 char* end = offset+len; | |
| 803 | |
| 804 for (char* eol; (cur<end) && (eol = find(cur, end, '\n')) < end; cur = eol + 1) | |
| 805 { | |
| 806 line = string(cur, eol); | |
| 807 if (line.back() == '\r') | |
| 808 line.pop_back(); | |
| 809 if (line.length() > 0) | |
| 810 { | |
| 811 if (line.at(0) == '>') | |
| 812 { | |
| 813 if (sequence.length() > KSIZE) | |
| 814 { // process previous sequence | |
| 815 target = process_read(sequence, acc, 0, sequence.length()-1); | |
| 816 } | |
| 817 sequence = ""; | |
| 818 acc = line.substr(1,line.length()-1); | |
| 819 } | |
| 820 else | |
| 821 { | |
| 822 sequence += line; | |
| 823 } | |
| 824 } | |
| 825 //cout << line << endl; | |
| 826 } | |
| 827 | |
| 828 // any trailing data in [eol, end) now is a partial line | |
| 829 offset = copy(cur, end, buf); | |
| 830 } | |
| 831 if (sequence.length() > KSIZE) | |
| 832 { //last one | |
| 833 target = process_read(sequence, acc, 0, sequence.length()-1); | |
| 834 } | |
| 835 // trailing data without eol | |
| 836 line = string(buf, offset); | |
| 837 if (gzclose(in) != Z_OK) error("failed gzclose"); | |
| 838 | |
| 839 } | |
| 840 | |
| 841 int process_gzkmers(gzFile in) | |
| 842 { | |
| 843 | |
| 844 char buf[BUFLEN]; | |
| 845 char* offset = buf; | |
| 846 string line; | |
| 847 int target, count, org, position, tct=0; | |
| 848 string sequence; | |
| 849 char strand; | |
| 850 bool fstrand; | |
| 851 | |
| 852 while (true) | |
| 853 { | |
| 854 int err, len = sizeof(buf)-(offset-buf); | |
| 855 if (len == 0) error("Buffer to small for input line lengths"); | |
| 856 | |
| 857 len = gzread(in, offset, len); | |
| 858 | |
| 859 if (len == 0) break; | |
| 860 if (len < 0) error(gzerror(in, &err)); | |
| 861 | |
| 862 char* cur = buf; | |
| 863 char* end = offset+len; | |
| 864 | |
| 865 for (char* eol; (cur<end) && (eol = find(cur, end, '\n')) < end; cur = eol + 1) | |
| 866 { | |
| 867 line = string(cur, eol); | |
| 868 if (line.back() == '\r') | |
| 869 line.pop_back(); | |
| 870 if (line.length() > 0) | |
| 871 { | |
| 872 replace(line.begin(), line.end(), ',', ' '); | |
| 873 istringstream ss(line); | |
| 874 if (ss >> sequence >> target >> org >> position >> strand >> count) | |
| 875 { | |
| 876 fstrand = (strand == 'F'); | |
| 877 process_kmer(sequence, target, org, position, fstrand); | |
| 878 tct++; | |
| 879 } | |
| 880 } | |
| 881 //cout << line << endl; | |
| 882 } | |
| 883 | |
| 884 // any trailing data in [eol, end) now is a partial line | |
| 885 offset = copy(cur, end, buf); | |
| 886 } | |
| 887 // trailing data without eol | |
| 888 line = string(buf, offset); | |
| 889 if (gzclose(in) != Z_OK) error("failed gzclose"); | |
| 890 | |
| 891 return tct; | |
| 892 | |
| 893 } | |
| 894 | |
| 895 void process_fq(string rname) | |
| 896 { //read fastq file and extract sequence and quality lines | |
| 897 //passes them to process_qual | |
| 898 string line, seq, acc= "", lseq; | |
| 899 ifstream fin; | |
| 900 int target; | |
| 901 int mod4=0; | |
| 902 | |
| 903 fin.open(rname); | |
| 904 if (fin) | |
| 905 { | |
| 906 while (getline(fin, line)) | |
| 907 { | |
| 908 if (line.back() == '\r') | |
| 909 line.pop_back(); | |
| 910 stringstream linestream(line); | |
| 911 linestream >> lseq; | |
| 912 if (lseq.length() > 0) | |
| 913 { | |
| 914 if (mod4 == 1) | |
| 915 { | |
| 916 seq = lseq; | |
| 917 } | |
| 918 else if (mod4 == 0) | |
| 919 { | |
| 920 acc = lseq; | |
| 921 } | |
| 922 else if (mod4 == 3) | |
| 923 { | |
| 924 process_qual(acc, seq, lseq); | |
| 925 } | |
| 926 mod4 = (mod4 + 1) % 4; | |
| 927 } | |
| 928 } | |
| 929 fin.close(); | |
| 930 } | |
| 931 } | |
| 932 | |
| 933 void process_fa(string rname) | |
| 934 { //read unzipped fasta file and extract sequence | |
| 935 //passes read to process_read | |
| 936 string sequence = "", line, lseq, acc= ""; | |
| 937 ifstream fin; | |
| 938 int target; | |
| 939 | |
| 940 fin.open(rname); | |
| 941 if (fin) | |
| 942 { | |
| 943 while (getline(fin, line)) | |
| 944 { | |
| 945 if (line.back() == '\r') | |
| 946 line.pop_back(); | |
| 947 stringstream linestream(line); | |
| 948 linestream >> lseq; | |
| 949 if (lseq[0] == '>') | |
| 950 { | |
| 951 if (sequence.length() > KSIZE) | |
| 952 { | |
| 953 target = process_read(sequence, acc, 0, sequence.length()-1); | |
| 954 } | |
| 955 sequence = ""; | |
| 956 acc = lseq.substr(1,lseq.length()-1); | |
| 957 } | |
| 958 else | |
| 959 sequence += lseq; | |
| 960 } | |
| 961 fin.close(); | |
| 962 if (sequence.length() > KSIZE) | |
| 963 { //last one | |
| 964 target = process_read(sequence, acc, 0, sequence.length()-1); | |
| 965 } | |
| 966 } | |
| 967 else | |
| 968 { | |
| 969 cout << "nark " << rname << endl; | |
| 970 } | |
| 971 } | |
| 972 | |
| 973 int main(int argc, char* argv[]) | |
| 974 { | |
| 975 int i,j,num_jobs; | |
| 976 vtype target; | |
| 977 ptype position; | |
| 978 char strand; | |
| 979 bool fstrand; | |
| 980 string header, jstr, fname, line, r1name, r2name = "", oname1, oname2; | |
| 981 int refi, targi, strni, count, org, num_targ; | |
| 982 ifstream fin; | |
| 983 string genus, species, acc, sequence, lseq, remainder; | |
| 984 vector<vector<string>> fnames; | |
| 985 vector<string> jnames; | |
| 986 vector<int> jcounts; | |
| 987 vector<string>::iterator it2; | |
| 988 string trname, s1, s2, s3,s4; | |
| 989 string argst, wdir = "", dname = "", jname = "", jfile = "", jdir = ""; | |
| 990 string iname = "mitochondria_data.txt"; //text file of groups for probe design | |
| 991 string tname = "mitochondria_tree.txt"; //edges in taxonomy tree | |
| 992 string pname = "mitochondria_probes.txt.gz"; | |
| 993 if (argc > 1) | |
| 994 { | |
| 995 for (i = 1; i < argc; i++) | |
| 996 { | |
| 997 argst = argv[i]; | |
| 998 if (argst == "-wdir") | |
| 999 { | |
| 1000 wdir = argv[i+1]; | |
| 1001 } | |
| 1002 if (argst == "-f1") | |
| 1003 { | |
| 1004 r1name =argv[i+1]; | |
| 1005 } | |
| 1006 if (argst == "-f2") | |
| 1007 { | |
| 1008 r2name =argv[i+1]; | |
| 1009 } | |
| 1010 } | |
| 1011 } | |
| 1012 iname = wdir + iname; | |
| 1013 tname = wdir + tname; | |
| 1014 pname = wdir + pname; | |
| 1015 | |
| 1016 ht = new Hashtable(); | |
| 1017 //load strain list | |
| 1018 | |
| 1019 | |
| 1020 cout << "r1 " << r1name << endl; | |
| 1021 cout << "r2 " << r2name << endl; | |
| 1022 cout << "wd " << wdir << endl; | |
| 1023 fin.open(iname); | |
| 1024 if (fin) | |
| 1025 { | |
| 1026 while (getline(fin, line)) | |
| 1027 { | |
| 1028 if (line.back() == '\r') | |
| 1029 line.pop_back(); | |
| 1030 if (line.length() > 1) | |
| 1031 { | |
| 1032 stringstream linestream(line); | |
| 1033 linestream >> targi >> acc; | |
| 1034 accession.push_back(acc); | |
| 1035 targno.push_back(targi); | |
| 1036 if (targi > num_targ) | |
| 1037 num_targ = targi; | |
| 1038 num_orgs++; | |
| 1039 } | |
| 1040 } | |
| 1041 fin.close(); | |
| 1042 cout << num_orgs << " strains" << endl; | |
| 1043 num_targ++; | |
| 1044 } | |
| 1045 taxonomy = new Tree1(num_targ); | |
| 1046 | |
| 1047 //load taxonomy tree | |
| 1048 fin.open(tname); | |
| 1049 if (fin) | |
| 1050 { | |
| 1051 while (getline(fin, line)) | |
| 1052 { | |
| 1053 if (line.back() == '\r') | |
| 1054 line.pop_back(); | |
| 1055 stringstream linestream(line); | |
| 1056 linestream >> i >> j; | |
| 1057 taxonomy->add_edge(i,j); | |
| 1058 } | |
| 1059 } | |
| 1060 else exit(1); | |
| 1061 fin.close(); | |
| 1062 cout << "tree loaded" << endl; | |
| 1063 | |
| 1064 //load kmer database | |
| 1065 tct = process_gzkmers(gzopen(pname.c_str(), "rb")); | |
| 1066 cout << tct << " kmers loaded" << endl; | |
| 1067 if (tct < 2) exit(1); | |
| 1068 | |
| 1069 //process reads | |
| 1070 gcount.resize(num_targ,0); | |
| 1071 ucount.resize(num_targ,0); | |
| 1072 fill(gcount.begin(), gcount.end(), 0); | |
| 1073 fill(ucount.begin(), ucount.end(), 0); | |
| 1074 kmer_seen.clear(); | |
| 1075 tct = 0; | |
| 1076 s1 = ".fastq.gz"; | |
| 1077 s2 = ".fasta"; | |
| 1078 s3 = ".fastq"; | |
| 1079 s4 = ".fasta.gz"; | |
| 1080 int slen = r1name.length(); | |
| 1081 char sch = r1name.at(slen-1); | |
| 1082 cout << slen << " : " << sch << endl; | |
| 1083 if (equal(s1.rbegin(), s1.rend(), r1name.rbegin())) | |
| 1084 { | |
| 1085 process_fqgz(gzopen(r1name.c_str(), "rb")); | |
| 1086 } | |
| 1087 else if (equal(s2.rbegin(), s2.rend(), r1name.rbegin())) | |
| 1088 { | |
| 1089 process_fa(r1name.c_str()); | |
| 1090 } | |
| 1091 else if (equal(s3.rbegin(), s3.rend(), r1name.rbegin())) | |
| 1092 { | |
| 1093 process_fq(r1name.c_str()); | |
| 1094 } | |
| 1095 else if (equal(s4.rbegin(), s4.rend(), r1name.rbegin())) | |
| 1096 { | |
| 1097 process_fagz(gzopen(r1name.c_str(), "rb")); | |
| 1098 } | |
| 1099 cout << tct << " reads loaded" << endl; | |
| 1100 if (r2name.length() > 1 && r2name != "none") | |
| 1101 { //second file? | |
| 1102 if (equal(s1.rbegin(), s1.rend(), r2name.rbegin())) | |
| 1103 { | |
| 1104 process_fqgz(gzopen(r2name.c_str(), "rb")); | |
| 1105 cout << tct << " reads loaded" << endl; | |
| 1106 } | |
| 1107 else if (equal(s2.rbegin(), s2.rend(), r2name.rbegin())) | |
| 1108 { | |
| 1109 process_fa(r2name.c_str()); | |
| 1110 } | |
| 1111 else if (equal(s3.rbegin(), s3.rend(), r2name.rbegin())) | |
| 1112 { | |
| 1113 process_fq(r2name.c_str()); | |
| 1114 } | |
| 1115 else if (equal(s4.rbegin(), s4.rend(), r2name.rbegin())) | |
| 1116 { | |
| 1117 process_fagz(gzopen(r2name.c_str(), "rb")); | |
| 1118 } | |
| 1119 cout << tct << " reads loaded" << endl; | |
| 1120 } | |
| 1121 | |
| 1122 oname2 = wdir + "result.txt"; | |
| 1123 ofstream out2(oname2); | |
| 1124 for (i=0; i<num_targ; i++) | |
| 1125 out2 << i << "," << gcount[i] << "," << ucount[i] << endl; | |
| 1126 out2.close(); | |
| 1127 | |
| 1128 delete taxonomy; | |
| 1129 delete ht; | |
| 1130 | |
| 1131 return 0; | |
| 1132 } |
