|
0
|
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 "fjoin.h"
|
|
|
13 #include "version.h"
|
|
|
14
|
|
|
15 using namespace std;
|
|
|
16
|
|
|
17 // define our program name
|
|
|
18 #define PROGRAM_NAME "fjoin"
|
|
|
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 forceStrand = false;
|
|
|
51 bool obeySplits = false;
|
|
|
52 bool inputIsBam = false;
|
|
|
53 bool outputIsBam = true;
|
|
|
54
|
|
|
55 // check to see if we should print out some help
|
|
|
56 if(argc <= 1) showHelp = true;
|
|
|
57
|
|
|
58 for(int i = 1; i < argc; i++) {
|
|
|
59 int parameterLength = (int)strlen(argv[i]);
|
|
|
60
|
|
|
61 if((PARAMETER_CHECK("-h", 2, parameterLength)) ||
|
|
|
62 (PARAMETER_CHECK("--help", 5, parameterLength))) {
|
|
|
63 showHelp = true;
|
|
|
64 }
|
|
|
65 }
|
|
|
66
|
|
|
67 if(showHelp) ShowHelp();
|
|
|
68
|
|
|
69 // do some parsing (all of these parameters require 2 strings)
|
|
|
70 for(int i = 1; i < argc; i++) {
|
|
|
71
|
|
|
72 int parameterLength = (int)strlen(argv[i]);
|
|
|
73
|
|
|
74 if(PARAMETER_CHECK("-a", 2, parameterLength)) {
|
|
|
75 if ((i+1) < argc) {
|
|
|
76 haveBedA = true;
|
|
|
77 outputIsBam = false;
|
|
|
78 bedAFile = argv[i + 1];
|
|
|
79 i++;
|
|
|
80 }
|
|
|
81 }
|
|
|
82 else if(PARAMETER_CHECK("-abam", 5, parameterLength)) {
|
|
|
83 if ((i+1) < argc) {
|
|
|
84 haveBedA = true;
|
|
|
85 inputIsBam = true;
|
|
|
86 bedAFile = argv[i + 1];
|
|
|
87 i++;
|
|
|
88 }
|
|
|
89 }
|
|
|
90 else if(PARAMETER_CHECK("-b", 2, parameterLength)) {
|
|
|
91 if ((i+1) < argc) {
|
|
|
92 haveBedB = true;
|
|
|
93 bedBFile = argv[i + 1];
|
|
|
94 i++;
|
|
|
95 }
|
|
|
96 }
|
|
|
97 else if(PARAMETER_CHECK("-bed", 4, parameterLength)) {
|
|
|
98 outputIsBam = false;
|
|
|
99 }
|
|
|
100 else if(PARAMETER_CHECK("-u", 2, parameterLength)) {
|
|
|
101 anyHit = true;
|
|
|
102 }
|
|
|
103 else if(PARAMETER_CHECK("-f", 2, parameterLength)) {
|
|
|
104 if ((i+1) < argc) {
|
|
|
105 haveFraction = true;
|
|
|
106 overlapFraction = atof(argv[i + 1]);
|
|
|
107 i++;
|
|
|
108 }
|
|
|
109 }
|
|
|
110 else if(PARAMETER_CHECK("-wa", 3, parameterLength)) {
|
|
|
111 writeA = true;
|
|
|
112 }
|
|
|
113 else if(PARAMETER_CHECK("-wb", 3, parameterLength)) {
|
|
|
114 writeB = true;
|
|
|
115 }
|
|
|
116 else if(PARAMETER_CHECK("-wo", 3, parameterLength)) {
|
|
|
117 writeOverlap = true;
|
|
|
118 }
|
|
|
119 else if(PARAMETER_CHECK("-wao", 4, parameterLength)) {
|
|
|
120 writeAllOverlap = true;
|
|
|
121 writeOverlap = true;
|
|
|
122 }
|
|
|
123 else if(PARAMETER_CHECK("-c", 2, parameterLength)) {
|
|
|
124 writeCount = true;
|
|
|
125 }
|
|
|
126 else if(PARAMETER_CHECK("-r", 2, parameterLength)) {
|
|
|
127 reciprocalFraction = true;
|
|
|
128 }
|
|
|
129 else if (PARAMETER_CHECK("-v", 2, parameterLength)) {
|
|
|
130 noHit = true;
|
|
|
131 }
|
|
|
132 else if (PARAMETER_CHECK("-s", 2, parameterLength)) {
|
|
|
133 forceStrand = true;
|
|
|
134 }
|
|
|
135 else if (PARAMETER_CHECK("-split", 6, parameterLength)) {
|
|
|
136 obeySplits = true;
|
|
|
137 }
|
|
|
138 else {
|
|
|
139 cerr << endl << "*****ERROR: Unrecognized parameter: " << argv[i] << " *****" << endl << endl;
|
|
|
140 showHelp = true;
|
|
|
141 }
|
|
|
142 }
|
|
|
143
|
|
|
144 // make sure we have both input files
|
|
|
145 if (!haveBedA || !haveBedB) {
|
|
|
146 cerr << endl << "*****" << endl << "*****ERROR: Need -a and -b files. " << endl << "*****" << endl;
|
|
|
147 showHelp = true;
|
|
|
148 }
|
|
|
149
|
|
|
150 if (anyHit && noHit) {
|
|
|
151 cerr << endl << "*****" << endl << "*****ERROR: Request either -u OR -v, not both." << endl << "*****" << endl;
|
|
|
152 showHelp = true;
|
|
|
153 }
|
|
|
154
|
|
|
155 if (writeB && writeCount) {
|
|
|
156 cerr << endl << "*****" << endl << "*****ERROR: Request either -wb OR -c, not both." << endl << "*****" << endl;
|
|
|
157 showHelp = true;
|
|
|
158 }
|
|
|
159
|
|
|
160 if (writeCount && writeOverlap) {
|
|
|
161 cerr << endl << "*****" << endl << "*****ERROR: Request either -wb OR -wo, not both." << endl << "*****" << endl;
|
|
|
162 showHelp = true;
|
|
|
163 }
|
|
|
164
|
|
|
165 if (writeA && writeOverlap) {
|
|
|
166 cerr << endl << "*****" << endl << "*****ERROR: Request either -wa OR -wo, not both." << endl << "*****" << endl;
|
|
|
167 showHelp = true;
|
|
|
168 }
|
|
|
169
|
|
|
170 if (writeB && writeOverlap) {
|
|
|
171 cerr << endl << "*****" << endl << "*****ERROR: Request either -wb OR -wo, not both." << endl << "*****" << endl;
|
|
|
172 showHelp = true;
|
|
|
173 }
|
|
|
174
|
|
|
175 if (reciprocalFraction && !haveFraction) {
|
|
|
176 cerr << endl << "*****" << endl << "*****ERROR: If using -r, you need to define -f." << endl << "*****" << endl;
|
|
|
177 showHelp = true;
|
|
|
178 }
|
|
|
179
|
|
|
180 if (anyHit && writeCount) {
|
|
|
181 cerr << endl << "*****" << endl << "*****ERROR: Request either -u OR -c, not both." << endl << "*****" << endl;
|
|
|
182 showHelp = true;
|
|
|
183 }
|
|
|
184
|
|
|
185 if (anyHit && writeB) {
|
|
|
186 cerr << endl << "*****" << endl << "*****ERROR: Request either -u OR -wb, not both." << endl << "*****" << endl;
|
|
|
187 showHelp = true;
|
|
|
188 }
|
|
|
189
|
|
|
190 if (anyHit && writeOverlap) {
|
|
|
191 cerr << endl << "*****" << endl << "*****ERROR: Request either -u OR -wo, not both." << endl << "*****" << endl;
|
|
|
192 showHelp = true;
|
|
|
193 }
|
|
|
194
|
|
|
195
|
|
|
196 if (!showHelp) {
|
|
|
197
|
|
|
198 BedIntersect *bi = new BedIntersect(bedAFile, bedBFile, anyHit, writeA, writeB, writeOverlap,
|
|
|
199 writeAllOverlap, overlapFraction, noHit, writeCount, forceStrand,
|
|
|
200 reciprocalFraction, obeySplits, inputIsBam, outputIsBam);
|
|
|
201 delete bi;
|
|
|
202 return 0;
|
|
|
203 }
|
|
|
204 else {
|
|
|
205 ShowHelp();
|
|
|
206 }
|
|
|
207 }
|
|
|
208
|
|
|
209 void ShowHelp(void) {
|
|
|
210
|
|
|
211 cerr << endl << "Program: " << PROGRAM_NAME << " (v" << VERSION << ")" << endl;
|
|
|
212
|
|
|
213 cerr << "Author: Aaron Quinlan (aaronquinlan@gmail.com)" << endl;
|
|
|
214
|
|
|
215 cerr << "Summary: Report overlaps between two feature files." << endl << endl;
|
|
|
216
|
|
|
217 cerr << "Usage: " << PROGRAM_NAME << " [OPTIONS] -a <bed/gff/vcf> -b <bed/gff/vcf>" << endl << endl;
|
|
|
218
|
|
|
219 cerr << "Options: " << endl;
|
|
|
220
|
|
|
221 cerr << "\t-abam\t" << "The A input file is in BAM format. Output will be BAM as well." << endl << endl;
|
|
|
222
|
|
|
223 cerr << "\t-bed\t" << "When using BAM input (-abam), write output as BED. The default" << endl;
|
|
|
224 cerr << "\t\tis to write output in BAM when using -abam." << endl << endl;
|
|
|
225
|
|
|
226 cerr << "\t-wa\t" << "Write the original entry in A for each overlap." << endl << endl;
|
|
|
227
|
|
|
228 cerr << "\t-wb\t" << "Write the original entry in B for each overlap." << endl;
|
|
|
229 cerr << "\t\t- Useful for knowing _what_ A overlaps. Restricted by -f and -r." << endl << endl;
|
|
|
230
|
|
|
231 cerr << "\t-wo\t" << "Write the original A and B entries plus the number of base" << endl;
|
|
|
232 cerr << "\t\tpairs of overlap between the two features." << endl;
|
|
|
233 cerr << "\t\t- Overlaps restricted by -f and -r." << endl;
|
|
|
234 cerr << "\t\t Only A features with overlap are reported." << endl << endl;
|
|
|
235
|
|
|
236 cerr << "\t-wao\t" << "Write the original A and B entries plus the number of base" << endl;
|
|
|
237 cerr << "\t\tpairs of overlap between the two features." << endl;
|
|
|
238 cerr << "\t\t- Overlapping features restricted by -f and -r." << endl;
|
|
|
239 cerr << "\t\t However, A features w/o overlap are also reported" << endl;
|
|
|
240 cerr << "\t\t with a NULL B feature and overlap = 0." << endl << endl;
|
|
|
241
|
|
|
242 cerr << "\t-u\t" << "Write the original A entry _once_ if _any_ overlaps found in B." << endl;
|
|
|
243 cerr << "\t\t- In other words, just report the fact >=1 hit was found." << endl;
|
|
|
244 cerr << "\t\t- Overlaps restricted by -f and -r." << endl << endl;
|
|
|
245
|
|
|
246 cerr << "\t-c\t" << "For each entry in A, report the number of overlaps with B." << endl;
|
|
|
247 cerr << "\t\t- Reports 0 for A entries that have no overlap with B." << endl;
|
|
|
248 cerr << "\t\t- Overlaps restricted by -f and -r." << endl << endl;
|
|
|
249
|
|
|
250 cerr << "\t-v\t" << "Only report those entries in A that have _no overlaps_ with B." << endl;
|
|
|
251 cerr << "\t\t- Similar to \"grep -v\" (an homage)." << endl << endl;
|
|
|
252
|
|
|
253 cerr << "\t-f\t" << "Minimum overlap required as a fraction of A." << endl;
|
|
|
254 cerr << "\t\t- Default is 1E-9 (i.e., 1bp)." << endl;
|
|
|
255 cerr << "\t\t- FLOAT (e.g. 0.50)" << endl << endl;
|
|
|
256
|
|
|
257 cerr << "\t-r\t" << "Require that the fraction overlap be reciprocal for A and B." << endl;
|
|
|
258 cerr << "\t\t- In other words, if -f is 0.90 and -r is used, this requires" << endl;
|
|
|
259 cerr << "\t\t that B overlap 90% of A and A _also_ overlaps 90% of B." << endl << endl;
|
|
|
260
|
|
|
261 cerr << "\t-s\t" << "Force strandedness. That is, only report hits in B that" << endl;
|
|
|
262 cerr << "\t\toverlap A on the same strand." << endl;
|
|
|
263 cerr << "\t\t- By default, overlaps are reported without respect to strand." << endl << endl;
|
|
|
264
|
|
|
265 cerr << "\t-split\t" << "Treat \"split\" BAM or BED12 entries as distinct BED intervals." << endl << endl;
|
|
|
266
|
|
|
267
|
|
|
268 // end the program here
|
|
|
269 exit(1);
|
|
|
270
|
|
|
271 }
|
