Mercurial > repos > aaronquinlan > multi_intersect

comparison BEDTools-Version-2.14.3/src/shuffleBed/shuffleBedMain.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 shuffleBedMain.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 "shuffleBed.h"
13 #include "version.h"
14
15 using namespace std;
16
17 // define our program name
18 #define PROGRAM_NAME "shuffleBed"
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 bedFile = "stdin";
34 string excludeFile;
35 string includeFile;
36 string genomeFile;
37
38 bool haveBed = true;
39 bool haveGenome = false;
40 bool haveExclude = false;
41 bool haveInclude = false;
42 bool haveSeed = false;
43 float overlapFraction = 0.0;
44 int seed = -1;
45 bool sameChrom = false;
46
47
48 for(int i = 1; i < argc; i++) {
49 int parameterLength = (int)strlen(argv[i]);
50
51 if((PARAMETER_CHECK("-h", 2, parameterLength)) ||
52 (PARAMETER_CHECK("--help", 5, parameterLength))) {
53 showHelp = true;
54 }
55 }
56
57 if(showHelp) ShowHelp();
58
59 // do some parsing (all of these parameters require 2 strings)
60 for(int i = 1; i < argc; i++) {
61
62 int parameterLength = (int)strlen(argv[i]);
63
64 if(PARAMETER_CHECK("-i", 2, parameterLength)) {
65 if ((i+1) < argc) {
66 bedFile = argv[i + 1];
67 i++;
68 }
69 }
70 else if(PARAMETER_CHECK("-g", 2, parameterLength)) {
71 if ((i+1) < argc) {
72 haveGenome = true;
73 genomeFile = argv[i + 1];
74 i++;
75 }
76 }
77 else if(PARAMETER_CHECK("-excl", 5, parameterLength)) {
78 if ((i+1) < argc) {
79 haveExclude = true;
80 excludeFile = argv[i + 1];
81 i++;
82 }
83 }
84 else if(PARAMETER_CHECK("-incl", 5, parameterLength)) {
85 if ((i+1) < argc) {
86 haveInclude = true;
87 includeFile = argv[i + 1];
88 i++;
89 }
90 }
91 else if(PARAMETER_CHECK("-seed", 5, parameterLength)) {
92 if ((i+1) < argc) {
93 haveSeed = true;
94 seed = atoi(argv[i + 1]);
95 i++;
96 }
97 }
98 else if(PARAMETER_CHECK("-chrom", 6, parameterLength)) {
99 sameChrom = true;
100 }
101 else if(PARAMETER_CHECK("-f", 2, parameterLength)) {
102 if ((i+1) < argc) {
103 overlapFraction = atof(argv[i + 1]);
104 i++;
105 }
106 }
107 else {
108 cerr << endl << "*****ERROR: Unrecognized parameter: " << argv[i] << " *****" << endl << endl;
109 showHelp = true;
110 }
111 }
112
113 // make sure we have both input files
114 if (!haveBed || !haveGenome) {
115 cerr << endl << "*****" << endl << "*****ERROR: Need both a BED (-i) and a genome (-g) file. " << endl << "*****" << endl;
116 showHelp = true;
117 }
118
119 if (haveInclude && haveExclude) {
120 cerr << endl << "*****" << endl << "*****ERROR: Cannot use -incl and -excl together." << endl << "*****" << endl;
121 showHelp = true;
122 }
123
124 if (!showHelp) {
125 BedShuffle *bc = new BedShuffle(bedFile, genomeFile, excludeFile, includeFile,
126 haveSeed, haveExclude, haveInclude, sameChrom,
127 overlapFraction, seed);
128 delete bc;
129 return 0;
130 }
131 else {
132 ShowHelp();
133 }
134 }
135
136 void ShowHelp(void) {
137
138 cerr << endl << "Program: " << PROGRAM_NAME << " (v" << VERSION << ")" << endl;
139
140 cerr << "Author: Aaron Quinlan (aaronquinlan@gmail.com)" << endl;
141
142 cerr << "Summary: Randomly permute the locations of a feature file among a genome." << endl << endl;
143
144 cerr << "Usage: " << PROGRAM_NAME << " [OPTIONS] -i <bed/gff/vcf> -g <genome>" << endl << endl;
145
146 cerr << "Options: " << endl;
147 cerr << "\t-excl\t" << "A BED/GFF/VCF file of coordinates in which features in -i" << endl;
148 cerr << "\t\tshould not be placed (e.g. gaps.bed)." << endl << endl;
149
150 cerr << "\t-incl\t" << "Instead of randomly placing features in a genome, the -incl" << endl;
151 cerr << "\t\toptions defines a BED/GFF/VCF file of coordinates in which " << endl;
152 cerr << "\t\tfeatures in -i should be randomly placed (e.g. genes.bed). " << endl << endl;
153
154 cerr << "\t-chrom\t" << "Keep features in -i on the same chromosome."<< endl;
155 cerr << "\t\t- By default, the chrom and position are randomly chosen." << endl << endl;
156
157 cerr << "\t-seed\t" << "Supply an integer seed for the shuffling." << endl;
158 cerr << "\t\t- By default, the seed is chosen automatically." << endl;
159 cerr << "\t\t- (INTEGER)" << endl << endl;
160
161 cerr << "\t-f\t" << "Maximum overlap (as a fraction of the -i feature) with an -excl" << endl;
162 cerr << "\t\tfeature that is tolerated before searching for a new, " << endl;
163 cerr << "\t\trandomized locus. For example, -f 0.10 allows up to 10%" << endl;
164 cerr << "\t\tof a randomized feature to overlap with a given feature" << endl;
165 cerr << "\t\tin the -excl file. **Cannot be used with -incl file.**" << endl;
166 cerr << "\t\t- Default is 1E-9 (i.e., 1bp)." << endl;
167 cerr << "\t\t- FLOAT (e.g. 0.50)" << endl << endl;
168
169 cerr << "Notes: " << endl;
170 cerr << "\t(1) The genome file should tab delimited and structured as follows:" << endl;
171 cerr << "\t <chromName><TAB><chromSize>" << endl << endl;
172 cerr << "\tFor example, Human (hg19):" << endl;
173 cerr << "\tchr1\t249250621" << endl;
174 cerr << "\tchr2\t243199373" << endl;
175 cerr << "\t..." << endl;
176 cerr << "\tchr18_gl000207_random\t4262" << endl << endl;
177
178
179 cerr << "Tips: " << endl;
180 cerr << "\tOne can use the UCSC Genome Browser's MySQL database to extract" << endl;
181 cerr << "\tchromosome sizes. For example, H. sapiens:" << endl << endl;
182 cerr << "\tmysql --user=genome --host=genome-mysql.cse.ucsc.edu -A -e \\" << endl;
183 cerr << "\t\"select chrom, size from hg19.chromInfo\" > hg19.genome" << endl << endl;
184
185
186 // end the program here
187 exit(1);
188 }