|
0
|
1 /*****************************************************************************
|
|
|
2 closestMain.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 "closestBed.h"
|
|
|
13 #include "version.h"
|
|
|
14
|
|
|
15 using namespace std;
|
|
|
16
|
|
|
17 // define our program name
|
|
|
18 #define PROGRAM_NAME "closestBed"
|
|
|
19
|
|
|
20 // define our parameter checking macro
|
|
|
21 #define PARAMETER_CHECK(param, paramLen, actualLen) (strncmp(argv[i], param, min(actualLen, paramLen))== 0) && (actualLen == paramLen)
|
|
|
22
|
|
|
23 // function declarations
|
|
|
24 void ShowHelp(void);
|
|
|
25
|
|
|
26 int main(int argc, char* argv[]) {
|
|
|
27
|
|
|
28 // our configuration variables
|
|
|
29 bool showHelp = false;
|
|
|
30
|
|
|
31 // input files
|
|
|
32 string bedAFile;
|
|
|
33 string bedBFile;
|
|
|
34 string tieMode = "all";
|
|
|
35 string strandedDistMode = "";
|
|
|
36
|
|
|
37 bool haveBedA = false;
|
|
|
38 bool haveBedB = false;
|
|
|
39 bool haveTieMode = false;
|
|
|
40 bool sameStrand = false;
|
|
|
41 bool diffStrand = false;
|
|
|
42 bool ignoreOverlaps = false;
|
|
|
43 bool reportDistance = false;
|
|
|
44 bool signDistance = false;
|
|
|
45 bool haveStrandedDistMode = false;
|
|
|
46
|
|
|
47
|
|
|
48 // check to see if we should print out some help
|
|
|
49 if(argc <= 1) showHelp = true;
|
|
|
50
|
|
|
51 for(int i = 1; i < argc; i++) {
|
|
|
52 int parameterLength = (int)strlen(argv[i]);
|
|
|
53
|
|
|
54 if( (PARAMETER_CHECK("-h", 2, parameterLength)) ||
|
|
|
55 (PARAMETER_CHECK("--help", 5, parameterLength))) {
|
|
|
56 showHelp = true;
|
|
|
57 }
|
|
|
58 }
|
|
|
59
|
|
|
60 if(showHelp) ShowHelp();
|
|
|
61
|
|
|
62 // do some parsing (all of these parameters require 2 strings)
|
|
|
63 for(int i = 1; i < argc; i++) {
|
|
|
64
|
|
|
65 int parameterLength = (int)strlen(argv[i]);
|
|
|
66
|
|
|
67 if(PARAMETER_CHECK("-a", 2, parameterLength)) {
|
|
|
68 if ((i+1) < argc) {
|
|
|
69 haveBedA = true;
|
|
|
70 bedAFile = argv[i + 1];
|
|
|
71 i++;
|
|
|
72 }
|
|
|
73 }
|
|
|
74 else if(PARAMETER_CHECK("-b", 2, parameterLength)) {
|
|
|
75 if ((i+1) < argc) {
|
|
|
76 haveBedB = true;
|
|
|
77 bedBFile = argv[i + 1];
|
|
|
78 i++;
|
|
|
79 }
|
|
|
80 }
|
|
|
81 else if (PARAMETER_CHECK("-s", 2, parameterLength)) {
|
|
|
82 sameStrand = true;
|
|
|
83 }
|
|
|
84 else if (PARAMETER_CHECK("-S", 2, parameterLength)) {
|
|
|
85 diffStrand = true;
|
|
|
86 }
|
|
|
87 else if (PARAMETER_CHECK("-d", 2, parameterLength)) {
|
|
|
88 reportDistance = true;
|
|
|
89 }
|
|
|
90 else if (PARAMETER_CHECK("-D", 2, parameterLength)) {
|
|
|
91 if ((i+1) < argc) {
|
|
|
92 reportDistance = true;
|
|
|
93 signDistance = true;
|
|
|
94 haveStrandedDistMode = true;
|
|
|
95 strandedDistMode = argv[i + 1];
|
|
|
96 i++;
|
|
|
97 }
|
|
|
98 }
|
|
|
99 else if (PARAMETER_CHECK("-io", 3, parameterLength)) {
|
|
|
100 ignoreOverlaps = true;
|
|
|
101 }
|
|
|
102 else if (PARAMETER_CHECK("-t", 2, parameterLength)) {
|
|
|
103 if ((i+1) < argc) {
|
|
|
104 haveTieMode = true;
|
|
|
105 tieMode = argv[i + 1];
|
|
|
106 i++;
|
|
|
107 }
|
|
|
108 }
|
|
|
109 else {
|
|
|
110 cerr << endl << "*****ERROR: Unrecognized parameter: " << argv[i] << " *****" << endl << endl;
|
|
|
111 showHelp = true;
|
|
|
112 }
|
|
|
113 }
|
|
|
114
|
|
|
115 // make sure we have both input files
|
|
|
116 if (!haveBedA || !haveBedB) {
|
|
|
117 cerr << endl << "*****" << endl << "*****ERROR: Need -a and -b files. " << endl << "*****" << endl;
|
|
|
118 showHelp = true;
|
|
|
119 }
|
|
|
120
|
|
|
121 if (haveTieMode && (tieMode != "all") && (tieMode != "first")
|
|
|
122 && (tieMode != "last")) {
|
|
|
123 cerr << endl << "*****" << endl << "*****ERROR: Request \"all\" or \"first\" or \"last\" for Tie Mode (-t)" << endl << "*****" << endl;
|
|
|
124 showHelp = true;
|
|
|
125 }
|
|
|
126
|
|
|
127 if (haveStrandedDistMode && (strandedDistMode != "a") && (strandedDistMode != "b")
|
|
|
128 && (strandedDistMode != "ref")) {
|
|
|
129 cerr << endl << "*****" << endl << "*****ERROR: Request \"a\" or \"b\" or \"ref\" for Stranded Distance Mode (-D)" << endl << "*****" << endl;
|
|
|
130 showHelp = true;
|
|
|
131 }
|
|
|
132
|
|
|
133 if (sameStrand && diffStrand) {
|
|
|
134 cerr << endl << "*****" << endl << "*****ERROR: Request either -s OR -S, not both." << endl << "*****" << endl;
|
|
|
135 showHelp = true;
|
|
|
136 }
|
|
|
137
|
|
|
138 if (!showHelp) {
|
|
|
139 BedClosest *bc = new BedClosest(bedAFile, bedBFile, sameStrand, diffStrand, tieMode, reportDistance, signDistance, strandedDistMode, ignoreOverlaps);
|
|
|
140 delete bc;
|
|
|
141 return 0;
|
|
|
142 }
|
|
|
143 else {
|
|
|
144 ShowHelp();
|
|
|
145 }
|
|
|
146 }
|
|
|
147
|
|
|
148 void ShowHelp(void) {
|
|
|
149
|
|
|
150 cerr << endl << "Program: " << PROGRAM_NAME << " (v" << VERSION << ")" << endl;
|
|
|
151
|
|
|
152 cerr << "Authors: Aaron Quinlan (aaronquinlan@gmail.com)" << endl;
|
|
|
153 cerr << "\t Erik Arner, Riken" << endl << endl;
|
|
|
154
|
|
|
155 cerr << "Summary: For each feature in A, finds the closest " << endl;
|
|
|
156 cerr << "\t feature (upstream or downstream) in B." << endl << endl;
|
|
|
157
|
|
|
158 cerr << "Usage: " << PROGRAM_NAME << " [OPTIONS] -a <bed/gff/vcf> -b <bed/gff/vcf>" << endl << endl;
|
|
|
159
|
|
|
160 cerr << "Options: " << endl;
|
|
|
161 cerr << "\t-s\t" << "Require same strandedness. That is, find the closest feature in B" << endl;
|
|
|
162 cerr << "\t\tthat overlaps A on the _same_ strand." << endl;
|
|
|
163 cerr << "\t\t- By default, overlaps are reported without respect to strand." << endl << endl;
|
|
|
164
|
|
|
165 cerr << "\t-S\t" << "Require opposite strandedness. That is, find the closest feature in B" << endl;
|
|
|
166 cerr << "\t\tthat overlaps A on the _opposite_ strand." << endl;
|
|
|
167 cerr << "\t\t- By default, overlaps are reported without respect to strand." << endl << endl;
|
|
|
168
|
|
|
169 cerr << "\t-d\t" << "In addition to the closest feature in B, " << endl;
|
|
|
170 cerr << "\t\treport its distance to A as an extra column." << endl;
|
|
|
171 cerr << "\t\t- The reported distance for overlapping features will be 0." << endl << endl;
|
|
|
172
|
|
|
173 cerr << "\t-D\t" << "Like -d, report the closest feature in B, and its distance to A" << endl;
|
|
|
174 cerr << "\t\tas an extra column. Unlike -d, use negative distances to report" << endl;
|
|
|
175 cerr << "\t\tupstream features. You must specify which orientation defines \"upstream\"." << endl;
|
|
|
176 cerr << "\t\tThe options are:" << endl;
|
|
|
177 cerr << "\t\t- \"ref\" Report distance with respect to the reference genome. " << endl;
|
|
|
178 cerr << "\t\t B features with a lower (start, stop) are upstream" << endl;
|
|
|
179 cerr << "\t\t- \"a\" Report distance with respect to A." << endl;
|
|
|
180 cerr << "\t\t When A is on the - strand, \"upstream\" means B has a higher (start,stop)." << endl;
|
|
|
181 cerr << "\t\t- \"b\" Report distance with respect to B." << endl;
|
|
|
182 cerr << "\t\t When B is on the - strand, \"upstream\" means A has a higher (start,stop)." << endl << endl;
|
|
|
183
|
|
|
184 cerr << "\t-io\t" << "Ignore features in B that overlap A. That is, we want close, but " << endl;
|
|
|
185 cerr << "\t\tnot touching features only." << endl << endl;
|
|
|
186
|
|
|
187 cerr << "\t-t\t" << "How ties for closest feature are handled. This occurs when two" << endl;
|
|
|
188 cerr << "\t\tfeatures in B have exactly the same \"closeness\" with A." << endl;
|
|
|
189 cerr << "\t\tBy default, all such features in B are reported." << endl;
|
|
|
190 cerr << "\t\tHere are all the options:" << endl;
|
|
|
191 cerr << "\t\t- \"all\" Report all ties (default)." << endl;
|
|
|
192 cerr << "\t\t- \"first\" Report the first tie that occurred in the B file." << endl;
|
|
|
193 cerr << "\t\t- \"last\" Report the last tie that occurred in the B file." << endl << endl;
|
|
|
194
|
|
|
195 cerr << "Notes: " << endl;
|
|
|
196 cerr << "\tReports \"none\" for chrom and \"-1\" for all other fields when a feature" << endl;
|
|
|
197 cerr << "\tis not found in B on the same chromosome as the feature in A." << endl;
|
|
|
198 cerr << "\tE.g. none\t-1\t-1" << endl << endl;
|
|
|
199
|
|
|
200 // end the program here
|
|
|
201 exit(1);
|
|
|
202 }
|
