|
0
|
1 /*****************************************************************************
|
|
|
2 multiBamCovMain.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 "multiBamCov.h"
|
|
|
13 #include "version.h"
|
|
|
14
|
|
|
15 using namespace std;
|
|
|
16
|
|
|
17 // define our program name
|
|
|
18 #define PROGRAM_NAME "multiBamCov"
|
|
|
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;
|
|
|
34 vector<string> bamFiles;
|
|
|
35 int minQual = 0;
|
|
|
36
|
|
|
37 // input arguments
|
|
|
38 bool haveBed = false;
|
|
|
39 bool haveBams = false;
|
|
|
40 bool properOnly = false;
|
|
|
41 bool keepDuplicates = false;
|
|
|
42 bool keepFailedQC = false;
|
|
|
43
|
|
|
44 // check to see if we should print out some help
|
|
|
45 if(argc <= 1) showHelp = true;
|
|
|
46
|
|
|
47 for(int i = 1; i < argc; i++) {
|
|
|
48 int parameterLength = (int)strlen(argv[i]);
|
|
|
49
|
|
|
50 if((PARAMETER_CHECK("-h", 2, parameterLength)) ||
|
|
|
51 (PARAMETER_CHECK("--help", 5, parameterLength))) {
|
|
|
52 showHelp = true;
|
|
|
53 }
|
|
|
54 }
|
|
|
55
|
|
|
56 if(showHelp) ShowHelp();
|
|
|
57
|
|
|
58 // do some parsing (all of these parameters require 2 strings)
|
|
|
59 for(int i = 1; i < argc; i++) {
|
|
|
60
|
|
|
61 int parameterLength = (int)strlen(argv[i]);
|
|
|
62
|
|
|
63 if(PARAMETER_CHECK("-bed", 4, parameterLength)) {
|
|
|
64 if ((i+1) < argc) {
|
|
|
65 haveBed = true;
|
|
|
66 bedFile = argv[i + 1];
|
|
|
67 i++;
|
|
|
68 }
|
|
|
69 }
|
|
|
70 else if(PARAMETER_CHECK("-bams", 5, parameterLength)) {
|
|
|
71 if ((i+1) < argc) {
|
|
|
72 haveBams = true;
|
|
|
73 i = i+1;
|
|
|
74 string file = argv[i];
|
|
|
75 while (file[0] != '-' && i < argc) {
|
|
|
76 bamFiles.push_back(file);
|
|
|
77 i++;
|
|
|
78 if (i < argc)
|
|
|
79 file = argv[i];
|
|
|
80 }
|
|
|
81 i--;
|
|
|
82 }
|
|
|
83 }
|
|
|
84 else if(PARAMETER_CHECK("-q", 2, parameterLength)) {
|
|
|
85 if ((i+1) < argc) {
|
|
|
86 minQual = atoi(argv[i + 1]);
|
|
|
87 i++;
|
|
|
88 }
|
|
|
89 }
|
|
|
90 else if(PARAMETER_CHECK("-p", 2, parameterLength)) {
|
|
|
91 properOnly = true;
|
|
|
92 }
|
|
|
93 else if(PARAMETER_CHECK("-D", 2, parameterLength)) {
|
|
|
94 keepDuplicates = true;
|
|
|
95 }
|
|
|
96
|
|
|
97 else if(PARAMETER_CHECK("-F", 2, parameterLength)) {
|
|
|
98 keepFailedQC = true;
|
|
|
99 }
|
|
|
100 else {
|
|
|
101 cerr << endl << "*****ERROR: Unrecognized parameter: " << argv[i] << " *****" << endl << endl;
|
|
|
102 showHelp = true;
|
|
|
103 }
|
|
|
104 }
|
|
|
105
|
|
|
106 if (!showHelp) {
|
|
|
107 MultiCovBam *mc = new MultiCovBam(bamFiles, bedFile, minQual, properOnly, keepDuplicates, keepFailedQC);
|
|
|
108 mc->CollectCoverage();
|
|
|
109 delete mc;
|
|
|
110 return 0;
|
|
|
111 }
|
|
|
112 else {
|
|
|
113 ShowHelp();
|
|
|
114 }
|
|
|
115 }
|
|
|
116
|
|
|
117 void ShowHelp(void) {
|
|
|
118
|
|
|
119 cerr << endl << "Program: " << PROGRAM_NAME << " (v" << VERSION << ")" << endl;
|
|
|
120
|
|
|
121 cerr << "Author: Aaron Quinlan (aaronquinlan@gmail.com)" << endl;
|
|
|
122
|
|
|
123 cerr << "Summary: Counts sequence coverage for multiple bams at specific loci." << endl << endl;
|
|
|
124
|
|
|
125 cerr << "Usage: " << PROGRAM_NAME << " [OPTIONS] -bams aln.1.bam aln.2.bam ... aln.n.bam -bed <bed/gff/vcf>" << endl << endl;
|
|
|
126
|
|
|
127 cerr << "Options: " << endl;
|
|
|
128
|
|
|
129 cerr << "\t-bams\t" << "The bam files." << endl << endl;
|
|
|
130
|
|
|
131 cerr << "\t-bed\t" << "The bed file." << endl << endl;
|
|
|
132
|
|
|
133 cerr << "\t-q\t" << "Minimum mapping quality allowed. Default is 0." << endl << endl;
|
|
|
134
|
|
|
135 cerr << "\t-D\t" << "Include duplicate-marked reads. Default is to count non-duplicates only" << endl << endl;
|
|
|
136
|
|
|
137 cerr << "\t-F\t" << "Include failed-QC reads. Default is to count pass-QC reads only" << endl << endl;
|
|
|
138
|
|
|
139 cerr << "\t-p\t" << "Only count proper pairs. Default is to count all alignments with MAPQ" << endl;
|
|
|
140 cerr << "\t\t" << "greater than the -q argument, regardless of the BAM FLAG field." << endl << endl;
|
|
|
141
|
|
|
142 // end the program here
|
|
|
143 exit(1);
|
|
|
144
|
|
|
145 }
|
