Mercurial > repos > aaronquinlan > multi_intersect

comparison BEDTools-Version-2.14.3/src/intersectBed/intersectMain.cpp @ 0:dfcd8b6c1bda

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Uploaded
author aaronquinlan
date Thu, 03 Nov 2011 10:25:04 -0400
parents
children
comparison
equal deleted inserted replaced
-1:000000000000 0:dfcd8b6c1bda
1 /*****************************************************************************
2 intersectMain.cpp
3
4 (c) 2009 - Aaron Quinlan
5 Hall Laboratory
6 Department of Biochemistry and Molecular Genetics
7 University of Virginia
8 aaronquinlan@gmail.com
9
10 Licenced under the GNU General Public License 2.0 license.
11 ******************************************************************************/
12 #include "intersectBed.h"
13 #include "version.h"
14
15 using namespace std;
16
17 // define our program name
18 #define PROGRAM_NAME "intersectBed"
19
20
21 // define our parameter checking macro
22 #define PARAMETER_CHECK(param, paramLen, actualLen) (strncmp(argv[i], param, min(actualLen, paramLen))== 0) && (actualLen == paramLen)
23
24 // function declarations
25 void ShowHelp(void);
26
27 int main(int argc, char* argv[]) {
28
29 // our configuration variables
30 bool showHelp = false;
31
32 // input files
33 string bedAFile;
34 string bedBFile;
35
36 // input arguments
37 float overlapFraction = 1E-9;
38
39 bool haveBedA = false;
40 bool haveBedB = false;
41 bool noHit = false;
42 bool anyHit = false;
43 bool writeA = false;
44 bool writeB = false;
45 bool writeCount = false;
46 bool writeOverlap = false;
47 bool writeAllOverlap = false;
48 bool haveFraction = false;
49 bool reciprocalFraction = false;
50 bool sameStrand = false;
51 bool diffStrand = false;
52 bool obeySplits = false;
53 bool inputIsBam = false;
54 bool outputIsBam = true;
55 bool uncompressedBam = false;
56 bool sortedInput = false;
57 // check to see if we should print out some help
58 if(argc <= 1) showHelp = true;
59
60 for(int i = 1; i < argc; i++) {
61 int parameterLength = (int)strlen(argv[i]);
62
63 if((PARAMETER_CHECK("-h", 2, parameterLength)) ||
64 (PARAMETER_CHECK("--help", 5, parameterLength))) {
65 showHelp = true;
66 }
67 }
68
69 if(showHelp) ShowHelp();
70
71 // do some parsing (all of these parameters require 2 strings)
72 for(int i = 1; i < argc; i++) {
73
74 int parameterLength = (int)strlen(argv[i]);
75
76 if(PARAMETER_CHECK("-a", 2, parameterLength)) {
77 if ((i+1) < argc) {
78 haveBedA = true;
79 outputIsBam = false;
80 bedAFile = argv[i + 1];
81 i++;
82 }
83 }
84 else if(PARAMETER_CHECK("-abam", 5, parameterLength)) {
85 if ((i+1) < argc) {
86 haveBedA = true;
87 inputIsBam = true;
88 bedAFile = argv[i + 1];
89 i++;
90 }
91 }
92 else if(PARAMETER_CHECK("-b", 2, parameterLength)) {
93 if ((i+1) < argc) {
94 haveBedB = true;
95 bedBFile = argv[i + 1];
96 i++;
97 }
98 }
99 else if(PARAMETER_CHECK("-bed", 4, parameterLength)) {
100 outputIsBam = false;
101 }
102 else if(PARAMETER_CHECK("-u", 2, parameterLength)) {
103 anyHit = true;
104 }
105 else if(PARAMETER_CHECK("-f", 2, parameterLength)) {
106 if ((i+1) < argc) {
107 haveFraction = true;
108 overlapFraction = atof(argv[i + 1]);
109 i++;
110 }
111 }
112 else if(PARAMETER_CHECK("-wa", 3, parameterLength)) {
113 writeA = true;
114 }
115 else if(PARAMETER_CHECK("-wb", 3, parameterLength)) {
116 writeB = true;
117 }
118 else if(PARAMETER_CHECK("-wo", 3, parameterLength)) {
119 writeOverlap = true;
120 }
121 else if(PARAMETER_CHECK("-wao", 4, parameterLength)) {
122 writeAllOverlap = true;
123 writeOverlap = true;
124 }
125 else if(PARAMETER_CHECK("-c", 2, parameterLength)) {
126 writeCount = true;
127 }
128 else if(PARAMETER_CHECK("-r", 2, parameterLength)) {
129 reciprocalFraction = true;
130 }
131 else if (PARAMETER_CHECK("-v", 2, parameterLength)) {
132 noHit = true;
133 }
134 else if (PARAMETER_CHECK("-s", 2, parameterLength)) {
135 sameStrand = true;
136 }
137 else if (PARAMETER_CHECK("-S", 2, parameterLength)) {
138 diffStrand = true;
139 }
140 else if (PARAMETER_CHECK("-split", 6, parameterLength)) {
141 obeySplits = true;
142 }
143 else if(PARAMETER_CHECK("-ubam", 5, parameterLength)) {
144 uncompressedBam = true;
145 }
146 else if(PARAMETER_CHECK("-sorted", 7, parameterLength)) {
147 sortedInput = true;
148 }
149 else {
150 cerr << endl << "*****ERROR: Unrecognized parameter: " << argv[i] << " *****" << endl << endl;
151 showHelp = true;
152 }
153 }
154
155 // make sure we have both input files
156 if (!haveBedA || !haveBedB) {
157 cerr << endl << "*****" << endl << "*****ERROR: Need -a and -b files. " << endl << "*****" << endl;
158 showHelp = true;
159 }
160
161 if (anyHit && noHit) {
162 cerr << endl << "*****" << endl << "*****ERROR: Request either -u OR -v, not both." << endl << "*****" << endl;
163 showHelp = true;
164 }
165
166 if (writeB && writeCount) {
167 cerr << endl << "*****" << endl << "*****ERROR: Request either -wb OR -c, not both." << endl << "*****" << endl;
168 showHelp = true;
169 }
170
171 if (writeCount && writeOverlap) {
172 cerr << endl << "*****" << endl << "*****ERROR: Request either -wb OR -wo, not both." << endl << "*****" << endl;
173 showHelp = true;
174 }
175
176 if (writeA && writeOverlap) {
177 cerr << endl << "*****" << endl << "*****ERROR: Request either -wa OR -wo, not both." << endl << "*****" << endl;
178 showHelp = true;
179 }
180
181 if (writeB && writeOverlap) {
182 cerr << endl << "*****" << endl << "*****ERROR: Request either -wb OR -wo, not both." << endl << "*****" << endl;
183 showHelp = true;
184 }
185
186 if (reciprocalFraction && !haveFraction) {
187 cerr << endl << "*****" << endl << "*****ERROR: If using -r, you need to define -f." << endl << "*****" << endl;
188 showHelp = true;
189 }
190
191 if (anyHit && writeCount) {
192 cerr << endl << "*****" << endl << "*****ERROR: Request either -u OR -c, not both." << endl << "*****" << endl;
193 showHelp = true;
194 }
195
196 if (anyHit && writeB) {
197 cerr << endl << "*****" << endl << "*****ERROR: Request either -u OR -wb, not both." << endl << "*****" << endl;
198 showHelp = true;
199 }
200
201 if (anyHit && writeOverlap) {
202 cerr << endl << "*****" << endl << "*****ERROR: Request either -u OR -wo, not both." << endl << "*****" << endl;
203 showHelp = true;
204 }
205
206 if (sameStrand && diffStrand) {
207 cerr << endl << "*****" << endl << "*****ERROR: Request either -s OR -S, not both." << endl << "*****" << endl;
208 showHelp = true;
209 }
210
211 if (!showHelp) {
212
213 BedIntersect *bi = new BedIntersect(bedAFile, bedBFile, anyHit, writeA, writeB, writeOverlap,
214 writeAllOverlap, overlapFraction, noHit, writeCount, sameStrand, diffStrand,
215 reciprocalFraction, obeySplits, inputIsBam, outputIsBam, uncompressedBam, sortedInput);
216 delete bi;
217 return 0;
218 }
219 else {
220 ShowHelp();
221 }
222 }
223
224 void ShowHelp(void) {
225
226 cerr << endl << "Program: " << PROGRAM_NAME << " (v" << VERSION << ")" << endl;
227
228 cerr << "Author: Aaron Quinlan (aaronquinlan@gmail.com)" << endl;
229
230 cerr << "Summary: Report overlaps between two feature files." << endl << endl;
231
232 cerr << "Usage: " << PROGRAM_NAME << " [OPTIONS] -a <bed/gff/vcf> -b <bed/gff/vcf>" << endl << endl;
233
234 cerr << "Options: " << endl;
235
236 cerr << "\t-abam\t" << "The A input file is in BAM format. Output will be BAM as well." << endl << endl;
237
238 cerr << "\t-ubam\t" << "Write uncompressed BAM output. Default is to write compressed BAM." << endl << endl;
239
240 cerr << "\t-bed\t" << "When using BAM input (-abam), write output as BED. The default" << endl;
241 cerr << "\t\tis to write output in BAM when using -abam." << endl << endl;
242
243 cerr << "\t-wa\t" << "Write the original entry in A for each overlap." << endl << endl;
244
245 cerr << "\t-wb\t" << "Write the original entry in B for each overlap." << endl;
246 cerr << "\t\t- Useful for knowing _what_ A overlaps. Restricted by -f and -r." << endl << endl;
247
248 cerr << "\t-wo\t" << "Write the original A and B entries plus the number of base" << endl;
249 cerr << "\t\tpairs of overlap between the two features." << endl;
250 cerr << "\t\t- Overlaps restricted by -f and -r." << endl;
251 cerr << "\t\t Only A features with overlap are reported." << endl << endl;
252
253 cerr << "\t-wao\t" << "Write the original A and B entries plus the number of base" << endl;
254 cerr << "\t\tpairs of overlap between the two features." << endl;
255 cerr << "\t\t- Overlapping features restricted by -f and -r." << endl;
256 cerr << "\t\t However, A features w/o overlap are also reported" << endl;
257 cerr << "\t\t with a NULL B feature and overlap = 0." << endl << endl;
258
259 cerr << "\t-u\t" << "Write the original A entry _once_ if _any_ overlaps found in B." << endl;
260 cerr << "\t\t- In other words, just report the fact >=1 hit was found." << endl;
261 cerr << "\t\t- Overlaps restricted by -f and -r." << endl << endl;
262
263 cerr << "\t-c\t" << "For each entry in A, report the number of overlaps with B." << endl;
264 cerr << "\t\t- Reports 0 for A entries that have no overlap with B." << endl;
265 cerr << "\t\t- Overlaps restricted by -f and -r." << endl << endl;
266
267 cerr << "\t-v\t" << "Only report those entries in A that have _no overlaps_ with B." << endl;
268 cerr << "\t\t- Similar to \"grep -v\" (an homage)." << endl << endl;
269
270 cerr << "\t-f\t" << "Minimum overlap required as a fraction of A." << endl;
271 cerr << "\t\t- Default is 1E-9 (i.e., 1bp)." << endl;
272 cerr << "\t\t- FLOAT (e.g. 0.50)" << endl << endl;
273
274 cerr << "\t-r\t" << "Require that the fraction overlap be reciprocal for A and B." << endl;
275 cerr << "\t\t- In other words, if -f is 0.90 and -r is used, this requires" << endl;
276 cerr << "\t\t that B overlap 90% of A and A _also_ overlaps 90% of B." << endl << endl;
277
278 cerr << "\t-s\t" << "Require same strandedness. That is, only report hits in B that" << endl;
279 cerr << "\t\toverlap A on the _same_ strand." << endl;
280 cerr << "\t\t- By default, overlaps are reported without respect to strand." << endl << endl;
281
282 cerr << "\t-S\t" << "Require different strandedness. That is, only report hits in B that" << endl;
283 cerr << "\t\toverlap A on the _opposite_ strand." << endl;
284 cerr << "\t\t- By default, overlaps are reported without respect to strand." << endl << endl;
285
286 cerr << "\t-split\t" << "Treat \"split\" BAM or BED12 entries as distinct BED intervals." << endl << endl;
287
288 cerr << "\t-sorted\t" << "Use the \"chromsweep\" algorithm for sorted (-k1,1 -k2,2n) input" << endl;
289 cerr << "\t\tNOTE: this will trust, but not enforce that data is sorted. Caveat emptor." << endl << endl;
290
291 // end the program here
292 exit(1);
293
294 }