|
0
|
1 /*
|
|
|
2 * Copyright (c) <2008 - 2012>, University of Washington, Simon Fraser University
|
|
|
3 * All rights reserved.
|
|
|
4 *
|
|
|
5 * Redistribution and use in source and binary forms, with or without modification,
|
|
|
6 * are permitted provided that the following conditions are met:
|
|
|
7 *
|
|
|
8 * Redistributions of source code must retain the above copyright notice, this list
|
|
|
9 * of conditions and the following disclaimer.
|
|
|
10 * - Redistributions in binary form must reproduce the above copyright notice, this
|
|
|
11 * list of conditions and the following disclaimer in the documentation and/or other
|
|
|
12 * materials provided with the distribution.
|
|
|
13 * - Neither the names of the University of Washington, Simon Fraser University,
|
|
|
14 * nor the names of its contributors may be
|
|
|
15 * used to endorse or promote products derived from this software without specific
|
|
|
16 * prior written permission.
|
|
|
17 *
|
|
|
18 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
|
|
|
19 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
|
|
|
20 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
|
|
|
21 * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
|
|
|
22 * CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
|
|
|
23 * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
|
|
|
24 * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
|
|
|
25 * PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
|
|
|
26 * LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
|
|
|
27 * NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
|
|
|
28 * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
|
|
|
29 */
|
|
|
30
|
|
|
31 /*
|
|
|
32 Authors:
|
|
|
33 Farhad Hormozdiari
|
|
|
34 Faraz Hach
|
|
|
35 Can Alkan
|
|
|
36 Emails:
|
|
|
37 farhadh AT uw DOT edu
|
|
|
38 fhach AT cs DOT sfu DOT ca
|
|
|
39 calkan AT uw DOT edu
|
|
|
40 */
|
|
|
41
|
|
|
42
|
|
|
43
|
|
|
44 #include <stdio.h>
|
|
|
45 #include <stdlib.h>
|
|
|
46 #include <string.h>
|
|
|
47 #include <math.h>
|
|
|
48 #include "Common.h"
|
|
|
49 #include "RefGenome.h"
|
|
|
50 #include "HashTable.h"
|
|
|
51 /**********************************************/
|
|
|
52 FILE *_ih_fp = NULL;
|
|
|
53 IHashTable *_ih_hashTable = NULL;
|
|
|
54 int _ih_maxHashTableSize = 0;
|
|
|
55 char *_ih_refGen = NULL;
|
|
|
56 char *_ih_refGenName = NULL;
|
|
|
57 long long _ih_memUsage = 0;
|
|
|
58 int _ih_refGenOff = 0;
|
|
|
59 /**********************************************/
|
|
|
60
|
|
|
61 int hashVal(char *seq)
|
|
|
62 {
|
|
|
63 int i=0;
|
|
|
64 int val=0, numericVal=0;
|
|
|
65
|
|
|
66 while(i<WINDOW_SIZE)
|
|
|
67 {
|
|
|
68 switch (seq[i])
|
|
|
69 {
|
|
|
70 case 'A':
|
|
|
71 numericVal = 0;
|
|
|
72 break;
|
|
|
73 case 'C':
|
|
|
74 numericVal = 1;
|
|
|
75 break;
|
|
|
76 case 'G' :
|
|
|
77 numericVal = 2;
|
|
|
78 break;
|
|
|
79 case 'T':
|
|
|
80 numericVal = 3;
|
|
|
81 break;
|
|
|
82 default:
|
|
|
83 return -1;
|
|
|
84 break;
|
|
|
85 }
|
|
|
86 val = (val << 2)|numericVal;
|
|
|
87 i++;
|
|
|
88 }
|
|
|
89 return val;
|
|
|
90 }
|
|
|
91 /**********************************************/
|
|
|
92 void freeIHashTableContent(IHashTable *hashTable, unsigned int maxSize)
|
|
|
93 {
|
|
|
94 int i=0;
|
|
|
95 for (i=0; i<maxSize; i++)
|
|
|
96 {
|
|
|
97 if (hashTable[i].locs != NULL)
|
|
|
98 {
|
|
|
99 freeMem(hashTable[i].locs, (hashTable[i].locs[0]+1)*(sizeof(unsigned int)));
|
|
|
100 hashTable[i].locs = NULL;
|
|
|
101 }
|
|
|
102 }
|
|
|
103 }
|
|
|
104 /**********************************************/
|
|
|
105 void initSavingIHashTable(char *fileName)
|
|
|
106 {
|
|
|
107 int tmp;
|
|
|
108 _ih_fp = fileOpen(fileName, "w");
|
|
|
109 unsigned char bIndex = 0; // Bisulfite Index
|
|
|
110
|
|
|
111 // First Two bytes are indicating the type of the index & window size
|
|
|
112 tmp = fwrite(&bIndex, sizeof(bIndex), 1, _ih_fp);
|
|
|
113 tmp = fwrite(&WINDOW_SIZE, sizeof(WINDOW_SIZE), 1, _ih_fp);
|
|
|
114
|
|
|
115 if (tmp == 0)
|
|
|
116 fprintf(stderr, "Write error while initializing hash table.\n");
|
|
|
117
|
|
|
118 }
|
|
|
119 /**********************************************/
|
|
|
120 void finalizeSavingIHashTable()
|
|
|
121 {
|
|
|
122 fclose(_ih_fp);
|
|
|
123 }
|
|
|
124 /**********************************************/
|
|
|
125 void saveIHashTable(IHashTable *hashTable, unsigned int size, unsigned int maxSize, char *refGen, char *refGenName, int refGenOffset)
|
|
|
126 {
|
|
|
127 int tmp;
|
|
|
128
|
|
|
129 // Every Chunk starts with a byte indicating whether it has extra info;
|
|
|
130 unsigned char extraInfo = 0;
|
|
|
131 tmp = fwrite (&extraInfo, sizeof(extraInfo), 1, _ih_fp);
|
|
|
132
|
|
|
133 short len = strlen(refGenName);
|
|
|
134 tmp = fwrite(&len, sizeof(len), 1, _ih_fp);
|
|
|
135 tmp = fwrite(refGenName, sizeof(char), len, _ih_fp);
|
|
|
136
|
|
|
137 tmp = fwrite(&refGenOffset, sizeof(refGenOffset), 1, _ih_fp);
|
|
|
138
|
|
|
139 unsigned int refGenLength = strlen(refGen);
|
|
|
140 tmp = fwrite(&refGenLength, sizeof(refGenLength), 1, _ih_fp);
|
|
|
141
|
|
|
142 tmp = fwrite(refGen, sizeof(char), refGenLength, _ih_fp);
|
|
|
143 tmp = fwrite(&size, sizeof(size), 1, _ih_fp);
|
|
|
144
|
|
|
145 int i=0,j=0;
|
|
|
146 unsigned char cnt=0;
|
|
|
147 for (i=0; i<maxSize; i++)
|
|
|
148 {
|
|
|
149 if (hashTable[i].locs != NULL)
|
|
|
150 {
|
|
|
151 tmp = fwrite(&i, sizeof(i), 1, _ih_fp);
|
|
|
152
|
|
|
153 if (hashTable[i].locs[0] < 250)
|
|
|
154 {
|
|
|
155 cnt = hashTable[i].locs[0];
|
|
|
156 tmp = fwrite(&cnt, sizeof(cnt), 1, _ih_fp);
|
|
|
157 }
|
|
|
158 else
|
|
|
159 {
|
|
|
160 cnt =0;
|
|
|
161 tmp = fwrite (&cnt, sizeof(cnt), 1, _ih_fp);
|
|
|
162 tmp = fwrite (&(hashTable[i].locs[0]), sizeof(hashTable[i].locs[0]), 1, _ih_fp);
|
|
|
163 }
|
|
|
164
|
|
|
165 for (j=1; j<=hashTable[i].locs[0]; j++)
|
|
|
166 tmp = fwrite(&(hashTable[i].locs[j]), sizeof(hashTable[i].locs[j]), 1, _ih_fp);
|
|
|
167 }
|
|
|
168 }
|
|
|
169
|
|
|
170 if (tmp == 0)
|
|
|
171 fprintf(stderr, "Write error while saving hash table.\n");
|
|
|
172 }
|
|
|
173 /**********************************************/
|
|
|
174 unsigned int addIHashTableLocation(IHashTable *hashTable, int hv, int location)
|
|
|
175 {
|
|
|
176 unsigned int sizeInc = 0;
|
|
|
177
|
|
|
178 if (hashTable[hv].locs == NULL)
|
|
|
179 {
|
|
|
180 sizeInc = 1;
|
|
|
181 hashTable[hv].locs = getMem (sizeof(unsigned int)*2);
|
|
|
182 hashTable[hv].locs[0]=1;
|
|
|
183 hashTable[hv].locs[1]=location;
|
|
|
184 }
|
|
|
185 else
|
|
|
186 {
|
|
|
187 int size = hashTable[hv].locs[0];
|
|
|
188 int i;
|
|
|
189 unsigned int *tmp = getMem( (size + 2) * sizeof(unsigned int) );
|
|
|
190
|
|
|
191 for (i = 0; i <= size; i++)
|
|
|
192 {
|
|
|
193 tmp[i] = hashTable[hv].locs[i];
|
|
|
194 }
|
|
|
195 size++;
|
|
|
196 tmp[0] = size;
|
|
|
197 tmp[size] = location;
|
|
|
198
|
|
|
199 freeMem(hashTable[hv].locs, (hashTable[hv].locs[0]*(sizeof(unsigned int))));
|
|
|
200 hashTable[hv].locs = tmp;
|
|
|
201 }
|
|
|
202 return sizeInc;
|
|
|
203 }
|
|
|
204 /**********************************************/
|
|
|
205 void generateIHashTable(char *fileName, char *indexName)
|
|
|
206 {
|
|
|
207 double startTime = getTime();
|
|
|
208 unsigned int hashTableSize = 0;
|
|
|
209 unsigned int hashTableMaxSize = pow(4, WINDOW_SIZE);
|
|
|
210 IHashTable *hashTable = getMem(sizeof(IHashTable)*hashTableMaxSize);
|
|
|
211 char *refGenName;
|
|
|
212 char *refGen;
|
|
|
213 int refGenOff = 0;
|
|
|
214 int i, hv, l, flag;
|
|
|
215
|
|
|
216
|
|
|
217 for ( i = 0; i < hashTableMaxSize; i++)
|
|
|
218 {
|
|
|
219 hashTable[i].locs = NULL;
|
|
|
220 }
|
|
|
221
|
|
|
222 //Loading Fasta File
|
|
|
223 if (!initLoadingRefGenome(fileName))
|
|
|
224 return;
|
|
|
225 initSavingIHashTable(indexName);
|
|
|
226
|
|
|
227 fprintf(stdout, "Generating Index from %s", fileName);
|
|
|
228 fflush(stdout);
|
|
|
229
|
|
|
230 char *prev = getMem (CONTIG_NAME_SIZE);
|
|
|
231 prev[0]='\0';
|
|
|
232
|
|
|
233 do
|
|
|
234 {
|
|
|
235 flag = loadRefGenome (&refGen, &refGenName, &refGenOff);
|
|
|
236
|
|
|
237 if ( strcmp(prev, refGenName) != 0)
|
|
|
238 {
|
|
|
239 fprintf(stdout, "\n - %s ", refGenName);
|
|
|
240 fflush(stdout);
|
|
|
241 sprintf(prev, "%s", refGenName);
|
|
|
242 }
|
|
|
243 else
|
|
|
244 {
|
|
|
245 fprintf(stdout, ".");
|
|
|
246 fflush(stdout);
|
|
|
247 }
|
|
|
248
|
|
|
249 l = strlen(refGen) - WINDOW_SIZE;
|
|
|
250
|
|
|
251 for (i=0; i < l; i++)
|
|
|
252 {
|
|
|
253 hv = hashVal(refGen + i);
|
|
|
254 if (hv != -1)
|
|
|
255 {
|
|
|
256 hashTableSize += addIHashTableLocation (hashTable, hv, i+1);
|
|
|
257 }
|
|
|
258 }
|
|
|
259
|
|
|
260 saveIHashTable(hashTable, hashTableSize, hashTableMaxSize, refGen, refGenName, refGenOff);
|
|
|
261 freeIHashTableContent(hashTable, hashTableMaxSize);
|
|
|
262 hashTableSize = 0;
|
|
|
263 } while (flag);
|
|
|
264
|
|
|
265 freeMem(prev, CONTIG_NAME_SIZE);
|
|
|
266 freeMem(hashTable, sizeof(IHashTable)*hashTableMaxSize);
|
|
|
267
|
|
|
268 finalizeLoadingRefGenome();
|
|
|
269 finalizeSavingIHashTable();
|
|
|
270
|
|
|
271 fprintf(stdout, "\nDONE in %0.2fs!\n", (getTime()-startTime));
|
|
|
272 }
|
|
|
273
|
|
|
274 /**********************************************/
|
|
|
275 void finalizeLoadingIHashTable()
|
|
|
276 {
|
|
|
277 freeIHashTableContent(_ih_hashTable, _ih_maxHashTableSize);
|
|
|
278 freeMem(_ih_hashTable, sizeof(IHashTable)* _ih_maxHashTableSize);
|
|
|
279 freeMem(_ih_refGen, strlen(_ih_refGen)+1) ;
|
|
|
280 freeMem(_ih_refGenName, strlen(_ih_refGenName)+1);
|
|
|
281 fclose(_ih_fp);
|
|
|
282 }
|
|
|
283
|
|
|
284 /**********************************************/
|
|
|
285 int loadIHashTable(double *loadTime, int errThreshould)
|
|
|
286 {
|
|
|
287 double startTime = getTime();
|
|
|
288 unsigned char extraInfo = 0;
|
|
|
289 short len;
|
|
|
290 unsigned int refGenLength;
|
|
|
291 unsigned int hashTableSize;
|
|
|
292 unsigned int tmpSize;
|
|
|
293 int tmp;
|
|
|
294 int i=0,j=0;
|
|
|
295
|
|
|
296 if ( fread(&extraInfo, sizeof(extraInfo), 1, _ih_fp) != sizeof(extraInfo) )
|
|
|
297 return 0;
|
|
|
298
|
|
|
299 freeIHashTableContent(_ih_hashTable, _ih_maxHashTableSize);
|
|
|
300 freeMem(_ih_refGen, strlen(_ih_refGen)+1) ;
|
|
|
301 freeMem(_ih_refGenName, strlen(_ih_refGenName)+1);
|
|
|
302
|
|
|
303 // Reading Chr Name
|
|
|
304 tmp = fread(&len, sizeof(len), 1, _ih_fp);
|
|
|
305 if (tmp == 0){
|
|
|
306 fprintf(stderr, "Read error while loading hash table.\n");
|
|
|
307 exit(0);
|
|
|
308 }
|
|
|
309
|
|
|
310 _ih_refGenName = getMem(sizeof(char)* (len+1));
|
|
|
311 tmp = fread(_ih_refGenName, sizeof(char), len, _ih_fp);
|
|
|
312 _ih_refGenName [len] ='\0';
|
|
|
313
|
|
|
314 tmp = fread(&_ih_refGenOff, sizeof (_ih_refGenOff), 1, _ih_fp);
|
|
|
315
|
|
|
316 // Reading Size and Content of Ref Genome
|
|
|
317 tmp = fread(&refGenLength, sizeof(refGenLength), 1, _ih_fp);
|
|
|
318 _ih_refGen = getMem(sizeof(char)*(refGenLength+1));
|
|
|
319 tmp = fread(_ih_refGen, sizeof(char), refGenLength, _ih_fp);
|
|
|
320 _ih_refGen[refGenLength]='\0';
|
|
|
321
|
|
|
322 //Reading Hashtable Size and Content
|
|
|
323 tmp = fread(&hashTableSize, sizeof(hashTableSize), 1, _ih_fp);
|
|
|
324
|
|
|
325 unsigned int hv;
|
|
|
326 unsigned char cnt=0;
|
|
|
327 unsigned int index = 1;
|
|
|
328 unsigned int tmpValue = 0;
|
|
|
329 for (i=0; i<hashTableSize; i++)
|
|
|
330 {
|
|
|
331 index = 1;
|
|
|
332 tmp = fread(&hv, sizeof(hv), 1, _ih_fp);
|
|
|
333 tmp = fread(&cnt, sizeof(cnt), 1, _ih_fp);
|
|
|
334
|
|
|
335 if (cnt>0)
|
|
|
336 {
|
|
|
337 tmpSize = cnt;
|
|
|
338 }
|
|
|
339 else
|
|
|
340 {
|
|
|
341 tmp = fread(&tmpSize, sizeof(tmpSize), 1, _ih_fp);
|
|
|
342 }
|
|
|
343
|
|
|
344 _ih_hashTable[hv].locs = getMem( sizeof(unsigned int)* (tmpSize+1) );
|
|
|
345 _ih_hashTable[hv].locs[0] = tmpSize;
|
|
|
346
|
|
|
347 /* for (j=1; j<=tmpSize; j++)
|
|
|
348 tmp = fread(&(_ih_hashTable[hv].locs[j]), sizeof(_ih_hashTable[hv].locs[j]), 1, _ih_fp);
|
|
|
349 */
|
|
|
350
|
|
|
351 if(tmpSize != 0)
|
|
|
352 {
|
|
|
353 tmp = fread(&(_ih_hashTable[hv].locs[index]), sizeof(_ih_hashTable[hv].locs[index]), 1, _ih_fp);
|
|
|
354 index++;
|
|
|
355 }
|
|
|
356
|
|
|
357 for (j=2; j<=tmpSize; j++)
|
|
|
358 {
|
|
|
359 tmp = fread(&tmpValue, sizeof(_ih_hashTable[hv].locs[index-1]), 1, _ih_fp);
|
|
|
360 if(abs(tmpValue-_ih_hashTable[hv].locs[index-1]) > errThreshould)
|
|
|
361 {
|
|
|
362 _ih_hashTable[hv].locs[index] = tmpValue;
|
|
|
363 index++;
|
|
|
364 }
|
|
|
365 }
|
|
|
366 _ih_hashTable[hv].locs[0] = index-1;
|
|
|
367 }
|
|
|
368 *loadTime = getTime()-startTime;
|
|
|
369 return 1;
|
|
|
370 }
|
|
|
371 /**********************************************/
|
|
|
372 unsigned int *getIHashTableCandidates(int hv)
|
|
|
373 {
|
|
|
374 if ( hv != -1 )
|
|
|
375 return _ih_hashTable[hv].locs;
|
|
|
376 else
|
|
|
377 return NULL;
|
|
|
378 }
|
|
|
379 /**********************************************/
|
|
|
380 /**********************************************/
|
|
|
381 /**********************************************/
|
|
|
382 void configHashTable()
|
|
|
383 {
|
|
|
384 if (WINDOW_SIZE <= 15)
|
|
|
385 {
|
|
|
386 generateHashTable = &generateIHashTable;
|
|
|
387 loadHashTable = &loadIHashTable;
|
|
|
388 finalizeLoadingHashTable = &finalizeLoadingIHashTable;
|
|
|
389 getCandidates = &getIHashTableCandidates;
|
|
|
390 }
|
|
|
391 else
|
|
|
392 {
|
|
|
393
|
|
|
394
|
|
|
395 }
|
|
|
396 }
|
|
|
397 /**********************************************/
|
|
|
398 int initLoadingHashTable(char *fileName)
|
|
|
399 {
|
|
|
400 int i;
|
|
|
401 unsigned char bsIndex;
|
|
|
402 int tmp;
|
|
|
403
|
|
|
404 _ih_fp = fileOpen(fileName, "r");
|
|
|
405
|
|
|
406 if (_ih_fp == NULL)
|
|
|
407 return 0;
|
|
|
408
|
|
|
409 tmp = fread(&bsIndex, sizeof(bsIndex), 1, _ih_fp);
|
|
|
410
|
|
|
411 if (tmp == 0){
|
|
|
412 fprintf(stderr, "Read error while inilializing hash table loading.\n");
|
|
|
413 exit(0);
|
|
|
414 }
|
|
|
415
|
|
|
416 if (bsIndex)
|
|
|
417 {
|
|
|
418 fprintf(stdout, "Error: Wrong Type of Index indicated");
|
|
|
419 return 0;
|
|
|
420 }
|
|
|
421
|
|
|
422 tmp = fread(&WINDOW_SIZE, sizeof(WINDOW_SIZE), 1, _ih_fp);
|
|
|
423
|
|
|
424 configHashTable();
|
|
|
425
|
|
|
426 if (_ih_maxHashTableSize != pow(4, WINDOW_SIZE))
|
|
|
427 {
|
|
|
428
|
|
|
429 if (_ih_hashTable != NULL)
|
|
|
430 {
|
|
|
431 freeIHashTableContent(_ih_hashTable, _ih_maxHashTableSize);
|
|
|
432 freeMem(_ih_hashTable, sizeof(IHashTable)* _ih_maxHashTableSize);
|
|
|
433 freeMem(_ih_refGen, strlen(_ih_refGen)+1) ;
|
|
|
434 freeMem(_ih_refGenName, strlen(_ih_refGenName)+1);
|
|
|
435 }
|
|
|
436
|
|
|
437 _ih_maxHashTableSize = pow(4, WINDOW_SIZE);
|
|
|
438
|
|
|
439 _ih_hashTable = getMem (sizeof(IHashTable) * _ih_maxHashTableSize);
|
|
|
440 for (i=0; i<_ih_maxHashTableSize; i++)
|
|
|
441 _ih_hashTable[i].locs = NULL;
|
|
|
442 _ih_refGen = getMem(1);
|
|
|
443 _ih_refGen[0]='\0';
|
|
|
444 _ih_refGenName = getMem(1);
|
|
|
445 _ih_refGenName[0] = '\0';
|
|
|
446 }
|
|
|
447
|
|
|
448 return 1;
|
|
|
449 }
|
|
|
450 /**********************************************/
|
|
|
451 char *getRefGenome()
|
|
|
452 {
|
|
|
453 return _ih_refGen;
|
|
|
454
|
|
|
455 }
|
|
|
456 /**********************************************/
|
|
|
457 char *getRefGenomeName()
|
|
|
458 {
|
|
|
459 return _ih_refGenName;
|
|
|
460
|
|
|
461 }
|
|
|
462 /**********************************************/
|
|
|
463 int getRefGenomeOffset()
|
|
|
464 {
|
|
|
465 return _ih_refGenOff;
|
|
|
466
|
|
|
467 }
|
|
|
468 /**********************************************/
|
|
|
469 HashTable *getHashTable()
|
|
|
470 {
|
|
|
471 return NULL;
|
|
|
472 }
|
