comparison deseq-hts_1.0/mex/get_reads.cpp @ 0:94a108763d9e draft

deseq-hts version 1.0 wraps the DESeq 1.6.0
author vipints
date Wed, 09 May 2012 20:43:47 -0400
parents
children
comparison
equal deleted inserted replaced
-1:000000000000 0:94a108763d9e
1 /*
2 * This program is free software; you can redistribute it and/or modify
3 * it under the terms of the GNU General Public License as published by
4 * the Free Software Foundation; either version 3 of the License, or
5 * (at your option) any later version.
6 *
7 * Written (W) 2010-2011 Jonas Behr, Regina Bohnert, Gunnar Raetsch
8 * Copyright (C) 2010-2011 Max Planck Society
9 */
10
11
12 #include <stdio.h>
13 #include <string.h>
14 #include <signal.h>
15 #include <mex.h>
16 #include <algorithm>
17 #include <vector>
18 using std::vector;
19 #include "get_reads_direct.h"
20 #include "mex_input.h"
21 #include "read.h"
22
23 #define MAXLINE 10000
24
25 /*
26 * input:
27 * 1 bam file
28 * 2 chromosome
29 * 3 region start (1-based index)
30 * 4 region end (1-based index)
31 * 5 strand (either '+' or '-' or '0')
32 * [6] collapse flag: if true the reads are collapsed to a coverage track
33 * [7] subsample percentage: percentage of reads to be subsampled (in per mill)
34 * [8] intron length filter
35 * [9] exon length filter
36 * [10] mismatch filter
37 * [11] bool: use mapped reads for coverage
38 * [12] bool: use spliced reads for coverage
39 * [13] return maxminlen
40 * [14] return pair coverage
41 *
42 * output:
43 * 1 coverage
44 * [2] intron cell array
45 * [3] pair coverage
46 * [4] pair list
47 *
48 * example call:
49 * [cov introns] = get_reads('polyA_left_I+_el15_mm1_spliced.bam', 'I', 10000, 12000, '-', 1, 30);
50 */
51 void mexFunction(int nlhs, mxArray *plhs[], int nrhs, const mxArray *prhs[]) {
52
53 if (nrhs<5 || nrhs>14 || (nlhs<1 || nlhs>4)) {
54 fprintf(stderr, "usage: [x [introns] [pair]] = get_reads(fname, chr, start, end, strand, [collapse], [subsample], [max intron length], [min exonlength], [max mismatches], [mapped], [spliced], [maxminlen], [pair]);\n");
55 return;
56 }
57
58 /* obligatory arguments
59 * **********************/
60 char *fname = get_string(prhs[0]);
61 //fprintf(stdout, "arg1: %s\n", fname);
62 char *chr = get_string(prhs[1]);
63 //fprintf(stdout, "arg2: %s\n", chr);
64 int from_pos = get_int(prhs[2]);
65 //fprintf(stdout, "arg3: %d\n", from_pos);
66 int to_pos = get_int(prhs[3]);
67 //fprintf(stdout, "arg4: %d\n", to_pos);
68 char *strand = get_string(prhs[4]);
69 //fprintf(stdout, "arg5: %s\n", strand);
70
71 if (from_pos>to_pos)
72 mexErrMsgTxt("Start (arg 3) must be <= end (arg 4)\n");
73
74 if (strand[0]!='+' && strand[0]!='-' && strand[0]!='0')
75 mexErrMsgTxt("Unknown strand (arg 5): either + or - or 0");
76
77 /* optional arguments
78 * ******************/
79 int collapse = 0;
80 if (nrhs>=6)
81 collapse = get_int(prhs[5]);
82
83 int subsample = 1000;
84 if (nrhs>=7)
85 subsample = get_int(prhs[6]);
86
87 int intron_len_filter = 1e9;
88 if (nrhs>=8)
89 intron_len_filter = get_int(prhs[7]);
90
91 int exon_len_filter = -1;
92 if (nrhs>=9)
93 exon_len_filter = get_int(prhs[8]);
94
95 int filter_mismatch = 1e9;
96 if (nrhs>=10)
97 filter_mismatch = get_int(prhs[9]);
98
99 int mapped = 1;
100 if (nrhs>=11)
101 mapped = get_int(prhs[10]);
102
103 int spliced = 1;
104 if (nrhs>=12)
105 spliced = get_int(prhs[11]);
106
107 int maxminlen = 0;
108 if (nrhs>=13)
109 maxminlen = get_int(prhs[12]);
110
111 int pair_cov = 0;
112 if (nrhs>=14)
113 pair_cov = get_int(prhs[13]);
114
115 /* call function to get reads
116 * **************************/
117 char region[MAXLINE];
118 sprintf(region, "%s:%i-%i", chr, from_pos, to_pos);
119
120 vector<CRead*> all_reads;
121
122 get_reads_from_bam(fname, region, &all_reads, strand[0], subsample);
123
124 /* filter reads
125 * **************/
126 int left = 0;
127 int right = 0;
128
129 vector<CRead*> reads;
130 for (int i=0; i<all_reads.size(); i++) {
131 if (all_reads[i]->left)
132 left++;
133 if (all_reads[i]->right)
134 right++;
135 if (all_reads[i]->max_intron_len()<intron_len_filter && all_reads[i]->min_exon_len()>exon_len_filter && all_reads[i]->get_mismatches()<=filter_mismatch)
136 reads.push_back(all_reads[i]);
137 }
138
139
140 /* prepare output
141 * **************/
142 int num_rows = reads.size();
143 int num_pos = to_pos-from_pos+1;
144
145 if (pair_cov==1 && nlhs>=3) {
146 // sort reads by read_id
147 printf("\n\nleft:%i right:%i \n\n", left, right);
148 sort(reads.begin(), reads.end(), CRead::compare_by_read_id);
149 }
150
151 // read coverages collapsed
152 if (collapse) {
153 plhs[0] = mxCreateNumericMatrix(1, num_pos, mxUINT32_CLASS, mxREAL);
154 uint32_t *mask_ret = (uint32_t*) mxGetData(plhs[0]);
155 if (num_pos>0 && mask_ret==NULL)
156 mexErrMsgTxt("Error allocating memory\n");
157 if (mapped && spliced) {
158 for (int i=0; i<reads.size(); i++) {
159 reads[i]->get_coverage(from_pos, to_pos, mask_ret);
160 }
161 } else {
162 for (int i=0; i<reads.size(); i++) {
163 ssize_t num_exons = reads[i]->block_starts.size();
164 if ((num_exons==1 && mapped) || (num_exons>1 && spliced))
165 reads[i]->get_coverage(from_pos, to_pos, mask_ret);
166 }
167 }
168 }
169 // reads not collapsed
170 else {
171 uint32_t nzmax = 0; // maximal number of nonzero elements
172 int len = to_pos-from_pos+1;
173 for (uint i=0; i<reads.size(); i++) {
174 for (uint n = 0; n < reads[i]->block_starts.size(); n++) {
175 uint32_t from, to;
176 if (reads[i]->block_starts[n]+reads[i]->start_pos-from_pos >= 0)
177 from = reads[i]->block_starts[n]+reads[i]->start_pos-from_pos;
178 else
179 from = 0;
180 if (reads[i]->block_starts[n]+reads[i]->start_pos-from_pos+reads[i]->block_lengths[n] >= 0)
181 to = reads[i]->block_starts[n]+reads[i]->start_pos-from_pos+reads[i]->block_lengths[n];
182 else
183 to = 0;
184 for (int bp=from; bp<to&bp<len; bp++) {
185 nzmax++;
186 }
187 }
188 }
189 // 1st row: row indices
190 // 2nd row: column indices
191 plhs[0] = mxCreateDoubleMatrix(2, nzmax, mxREAL);
192 double *mask_ret = (double*) mxGetData(plhs[0]);
193 if (nzmax>0 && mask_ret==NULL)
194 mexErrMsgTxt("Error allocating memory\n");
195 uint32_t mask_ret_c = 0; // counter
196 for (uint i=0; i<reads.size(); i++) {
197 reads[i]->get_reads_sparse(from_pos, to_pos, mask_ret, mask_ret_c, i);
198 }
199 if (mask_ret_c!=2*nzmax)
200 mexErrMsgTxt("Error filling index arrays for sparse matrix\n");
201 }
202 // introns
203 if (maxminlen==0 && nlhs>=2) {
204 vector<int> intron_list;
205 for (int i=0; i<reads.size(); i++) {
206 reads[i]->get_introns(&intron_list);
207 }
208
209 plhs[1] = mxCreateNumericMatrix(2, intron_list.size()/2, mxUINT32_CLASS, mxREAL);
210 uint32_t *p_intron_list = (uint32_t*) mxGetData(plhs[1]);
211 for (int p = 0; p<intron_list.size(); p++) {
212 p_intron_list[p] = intron_list[p];
213 }
214 intron_list.clear();
215 } else if (nlhs>=2) {
216 vector<uint32_t> intron_starts;
217 vector<uint32_t> intron_ends;
218 vector<uint32_t> block_len1;
219 vector<uint32_t> block_len2;
220 for (int i=0; i<reads.size(); i++) {
221 reads[i]->get_introns(&intron_starts, &intron_ends, &block_len1, &block_len2);
222 }
223
224 plhs[1] = mxCreateNumericMatrix(4, intron_starts.size(), mxINT32_CLASS, mxREAL);
225 uint32_t *p_intron_list = (uint32_t*) mxGetData(plhs[1]);
226 for (int p = 0; p<intron_starts.size(); p++) {
227 p_intron_list[4*p] = intron_starts[p];
228 p_intron_list[(4*p)+1] = intron_ends[p];
229 p_intron_list[(4*p)+2] = block_len1[p];
230 p_intron_list[(4*p)+3] = block_len2[p];
231 }
232 intron_starts.clear() ;
233 intron_ends.clear() ;
234 block_len1.clear() ;
235 block_len2.clear() ;
236 }
237 if (pair_cov==1 && nlhs>=3) {
238 plhs[2] = mxCreateNumericMatrix(1, num_pos, mxUINT32_CLASS, mxREAL);
239 uint32_t *p_pair_map = (uint32_t*) mxGetData(plhs[2]);
240 if (num_pos>0 && p_pair_map==NULL)
241 mexErrMsgTxt("Error allocating memory\n");
242
243 vector<int> pair_ids;
244
245 int take_cnt = 0;
246 int discard_cnt = 0;
247 // find consecutive reads with the same id
248 for (int i=0; i<((int) reads.size())-1; i++) {
249 int j = i+1;
250 while(j<reads.size() && strcmp(reads[i]->read_id, reads[j]->read_id) == 0) {
251 if ((reads[i]->left && reads[j]->right) || (reads[j]->left && reads[i]->right) && (reads[i]->reverse != reads[j]->reverse)) {
252 if (reads[i]->get_last_position()==-1 || reads[j]->get_last_position()==-1)
253 break;
254 if (reads[i]->get_last_position()<reads[j]->start_pos && reads[j]->start_pos-reads[i]->get_last_position()<60000) {
255 int from = std::max(0, reads[i]->get_last_position()-from_pos);
256 int to = std::min(num_pos-1, reads[j]->start_pos-from_pos);
257 pair_ids.push_back(i);
258 pair_ids.push_back(j);
259 for (int k=from; k<to; k++)
260 p_pair_map[k]++;
261 take_cnt++;
262 } else if (reads[i]->start_pos>reads[j]->get_last_position() && reads[j]->get_last_position()-reads[i]->start_pos<60000) {
263 int from = std::max(0, reads[j]->get_last_position()-from_pos);
264 int to = std::min(num_pos-1, reads[i]->start_pos-from_pos);
265 pair_ids.push_back(i);
266 pair_ids.push_back(j);
267 for (int k=from; k<to; k++)
268 p_pair_map[k]++;
269 take_cnt++;
270 } else
271 discard_cnt++;
272 }
273 else
274 discard_cnt++;
275 j++;
276 }
277 }
278 printf("take:%i, discard:%i \n", take_cnt, discard_cnt);
279
280 if (nlhs>=4) {
281 plhs[3] = mxCreateNumericMatrix(2, pair_ids.size()/2, mxUINT32_CLASS, mxREAL);
282 uint32_t *pair_ids_ret = (uint32_t*) mxGetData(plhs[3]);
283 if (pair_ids.size()>0 && pair_ids_ret==NULL)
284 mexErrMsgTxt("Error allocating memory\n");
285 for (int i=0; i<pair_ids.size(); i++) {
286 pair_ids_ret[i] = pair_ids[i];
287 }
288 }
289 }
290 for (int i=0; i<all_reads.size(); i++)
291 delete all_reads[i];
292 }
293