diff --git a/bin/closestBed b/bin/closestBed
index 7d70fa4c9c24981be21fa2db2743e4b3d8fdc37a..bfea0660a86324e852bdeaa3e77e9ca450659810 100755
Binary files a/bin/closestBed and b/bin/closestBed differ
diff --git a/bin/intersectBed b/bin/intersectBed
index d1728c9b3a0bc452860325bf8f93bdeab2ccf251..7da0bb606c63bb04bda4aef4c76d767edb703465 100755
Binary files a/bin/intersectBed and b/bin/intersectBed differ
diff --git a/bin/subtractBed b/bin/subtractBed
index b9aef6a5a1511ae1277fbd8060e929853dcfef5c..f9923e440b76272f5ed743ffdb388b6660f5ec37 100755
Binary files a/bin/subtractBed and b/bin/subtractBed differ
diff --git a/bin/windowBed b/bin/windowBed
index 981f606a0c2da844f596fef988433f1d5be6d788..85ce50d8d4374f475a00e5830a4908ce3e64e6ff 100755
Binary files a/bin/windowBed and b/bin/windowBed differ
diff --git a/obj/closestMain.o b/obj/closestMain.o
index 149f4e12e9000ffb85b92e1f27d1e56373009284..70cb6534b8929c342a92681fab835fcea1285dd3 100644
Binary files a/obj/closestMain.o and b/obj/closestMain.o differ
diff --git a/obj/intersectMain.o b/obj/intersectMain.o
index 77731f5f8f7989950a3d5202a0d1c62706ee1f02..a383639beb1a4c921d62be67f72febb439425144 100644
Binary files a/obj/intersectMain.o and b/obj/intersectMain.o differ
diff --git a/obj/subtractBed.o b/obj/subtractBed.o
index a0f04779c12993910a5b9419bad86d770d96ef28..825a85f1a5471f636a7d587a3caaa3f3a899204f 100644
Binary files a/obj/subtractBed.o and b/obj/subtractBed.o differ
diff --git a/obj/subtractMain.o b/obj/subtractMain.o
index 97cb5e145fb7f9caa9af450d0fcd181b5594c595..b68c1657d85bbdc57e3560bcd028ba937598bd69 100644
Binary files a/obj/subtractMain.o and b/obj/subtractMain.o differ
diff --git a/obj/windowMain.o b/obj/windowMain.o
index 215e18c15eff7a59d1fcef4bb36c7ef9df18e722..db72bf73134d0d182c9208316a97f7c81ce7a292 100644
Binary files a/obj/windowMain.o and b/obj/windowMain.o differ
diff --git a/src/closestBed/closestMain.cpp b/src/closestBed/closestMain.cpp
index bb30410d2ba62f6506d4e68a567d1eeed32efb1c..e8698eca5cbf89f337c212d2c1444d23dcc09465 100755
--- a/src/closestBed/closestMain.cpp
+++ b/src/closestBed/closestMain.cpp
@@ -108,22 +108,22 @@ void ShowHelp(void) {
cerr << " Hall Laboratory, University of Virginia" << endl;
cerr << "===============================================" << endl << endl;
cerr << "SUMMARY: For each feature in BED A, finds the closest " << endl;
- cerr << " feature (upstream or downstream) in BED B." << endl << endl;
+ cerr << "\t feature (upstream or downstream) in BED B." << endl << endl;
- cerr << "USAGE: " << PROGRAM_NAME << " [OPTIONS] -a <a.bed> -b <b.bed>" << endl << endl;
+ cerr << "USAGE: " << PROGRAM_NAME << " [OPTIONS] -a <a.bed> -b <b.bed>" << endl << endl;
cerr << "OPTIONS: " << endl;
cerr << " " << "-s\t" << "Force strandedness. That is, find the closest feature in B that" << endl;
cerr << "\toverlaps A on the same strand." << endl;
- cerr << "\t - By default, overlaps are reported without respect to strand." << endl << endl;
+ cerr << "\t- By default, overlaps are reported without respect to strand." << endl << endl;
cerr << " " << "-t\t" << "How ties for closest feature are handled. This occurs when two" << endl;
cerr << "\tfeatures in B have exactly the same overlap with a feature in A." << endl;
cerr << "\tBy default, all such features in B are reported." << endl;
cerr << "\tHere are all the options:" << endl;
- cerr << "\t - \"all\" Report all ties (default)." << endl;
- cerr << "\t - \"first\" Report the first tie that occurred in the B file." << endl;
- cerr << "\t - \"last\" Report the last tie that occurred in the B file." << endl << endl;
+ cerr << "\t- \"all\" Report all ties (default)." << endl;
+ cerr << "\t- \"first\" Report the first tie that occurred in the B file." << endl;
+ cerr << "\t- \"last\" Report the last tie that occurred in the B file." << endl << endl;
cerr << "NOTES: " << endl;
diff --git a/src/intersectBed/intersectMain.cpp b/src/intersectBed/intersectMain.cpp
index e0c657dbcd5ed366e389bd82f59b53c35289a583..64e5caef33e05dcea733bbed6cad819017307095 100755
--- a/src/intersectBed/intersectMain.cpp
+++ b/src/intersectBed/intersectMain.cpp
@@ -158,34 +158,34 @@ void ShowHelp(void) {
cerr << "===============================================" << endl << endl;
cerr << "SUMMARY: Report overlaps between a.bed and b.bed." << endl << endl;
- cerr << "USAGE: " << PROGRAM_NAME << " [OPTIONS] -a <a.bed> -b <b.bed>" << endl << endl;
+ cerr << "USAGE: " << PROGRAM_NAME << " [OPTIONS] -a <a.bed> -b <b.bed>" << endl << endl;
cerr << "OPTIONS: " << endl;
cerr << " " << "-wa\t" << "Write the original entry in A for each overlap." << endl << endl;
cerr << " " << "-wb\t" << "Write the original entry in B for each overlap." << endl;
- cerr << "\t - Useful for knowing _what_ A overlaps. Restricted by -f." << endl << endl;
+ cerr << "\t- Useful for knowing _what_ A overlaps. Restricted by -f." << endl << endl;
cerr << " " << "-u\t" << "Write the original A entry _once_ if _any_ overlaps found in B." << endl;
- cerr << "\t - In other words, just report the fact >=1 hit was found." << endl << endl;
+ cerr << "\t- In other words, just report the fact >=1 hit was found." << endl << endl;
cerr << " " << "-c\t" << "For each entry in A, report the number of overlaps with B." << endl;
- cerr << "\t - Reports 0 for A entries that have no overlap with B." << endl;
- cerr << "\t - Overlaps restricted by -f." << endl << endl;
+ cerr << "\t- Reports 0 for A entries that have no overlap with B." << endl;
+ cerr << "\t- Overlaps restricted by -f." << endl << endl;
cerr << " " << "-v\t" << "Only report those entries in A that have _no overlaps_ with B." << endl;
- cerr << "\t - Similar to \"grep -v.\"" << endl << endl;
+ cerr << "\t- Similar to \"grep -v.\"" << endl << endl;
cerr << " " << "-f\t" << "Minimum overlap required as a fraction of A." << endl;
- cerr << "\t - Default is 1E-9 (i.e., 1bp)." << endl << endl;
+ cerr << "\t- Default is 1E-9 (i.e., 1bp)." << endl << endl;
cerr << " " << "-r\t" << "Require that the fraction overlap be reciprocal for A and B." << endl;
- cerr << "\t - In other words, if -f is 0.90 and -r is used, this requires" << endl;
- cerr << "\t that B overlap 90% of A and A _also_ overlaps 90% of B." << endl << endl;
+ cerr << "\t- In other words, if -f is 0.90 and -r is used, this requires" << endl;
+ cerr << "\t that B overlap 90% of A and A _also_ overlaps 90% of B." << endl << endl;
cerr << " " << "-s\t" << "Force strandedness. That is, only report hits in B that" << endl;
cerr << "\toverlap A on the same strand." << endl;
- cerr << "\t - By default, overlaps are reported without respect to strand." << endl << endl;
+ cerr << "\t- By default, overlaps are reported without respect to strand." << endl << endl;
// end the program here
exit(1);
diff --git a/src/subtractBed/b.bed b/src/subtractBed/b.bed
index 68a8ea6259d8e67e231439fc47ae525dc1cbf064..bfd157e3360d38f823f639f075a7c4aef6481b9c 100644
--- a/src/subtractBed/b.bed
+++ b/src/subtractBed/b.bed
@@ -1,2 +1,2 @@
-chr1 150 250
+chr1 180 300
diff --git a/src/subtractBed/subtractBed.cpp b/src/subtractBed/subtractBed.cpp
index 186c12ddf7c17a8e5685a480ab04889a75f700f7..7f22f6738021f4a9aab4360dad95280f9371c5ca 100755
--- a/src/subtractBed/subtractBed.cpp
+++ b/src/subtractBed/subtractBed.cpp
@@ -1,12 +1,14 @@
-//
-// subtractBed.cpp
-// BEDTools
-//
-// Created by Aaron Quinlan Spring 2009.
-// Copyright 2009 Aaron Quinlan. All rights reserved.
-//
-// Summary: Removes overlapping segments from a BED entry.
-//
+/*****************************************************************************
+ subtractBed.cpp
+
+ (c) 2009 - Aaron Quinlan
+ Hall Laboratory
+ Department of Biochemistry and Molecular Genetics
+ University of Virginia
+ aaronquinlan@gmail.com
+
+ Licenced under the GNU General Public License 2.0+ license.
+******************************************************************************/
#include "lineFileUtilities.h"
#include "subtractBed.h"
@@ -60,7 +62,7 @@ void BedSubtract::FindOverlaps(BED &a, vector<BED> &hits) {
// is there enough overlap (default ~ 1bp)
float overlap = ((float)(e-s) / (float)(a.end - a.start));
-
+
if (overlap >= 1.0) {
numOverlaps++;
numConsumedByB++;
@@ -147,74 +149,52 @@ void BedSubtract::FindOverlaps(BED &a, vector<BED> &hits) {
-void BedSubtract::SubtractBed() {
+void BedSubtract::SubtractBed(istream &bedInput) {
// load the "B" bed file into a map so
// that we can easily compare "A" to it for overlaps
bedB->loadBedFileIntoMap();
- string bedLine;
- BED bedEntry;
- int lineNum = 0;
-
- // are we dealing with a file?
- if (bedA->bedFile != "stdin") {
-
- // open the BED file for reading
- ifstream bed(bedA->bedFile.c_str(), ios::in);
- if ( !bed ) {
- cerr << "Error: The requested bed file (" <<bedA->bedFile << ") could not be opened. Exiting!" << endl;
- exit (1);
- }
+ string bedLine;
+ int lineNum = 0; // current input line number
+ vector<BED> hits; // vector of potential hits
+ vector<string> bedFields; // vector for a BED entry
+
+ // reserve some space
+ hits.reserve(100);
+ bedFields.reserve(12);
- BED a;
- // process each entry in A
- while (getline(bed, bedLine)) {
-
- if ((bedLine.find("track") != string::npos) || (bedLine.find("browser") != string::npos)) {
- continue;
- }
- else {
- // split the current line into ditinct fields
- vector<string> bedFields;
- Tokenize(bedLine,bedFields);
+ BED a;
+ // process each entry in A
+ while (getline(bedInput, bedLine)) {
- lineNum++;
-
- // find the overlaps with B if it's a valid BED entry.
- if (bedA->parseBedLine(a, bedFields, lineNum)) {
- vector<BED> hits;
- FindOverlaps(a, hits);
- }
- }
+ lineNum++;
+ Tokenize(bedLine,bedFields);
+
+ // find the overlaps with B if it's a valid BED entry.
+ if (bedA->parseLine(a, bedFields, lineNum)) {
+ FindOverlaps(a, hits);
+ hits.clear();
}
- }
- // "A" is being passed via STDIN.
- else {
- BED a;
- // process each entry in A
- while (getline(cin, bedLine)) {
+ // reset for the next input line
+ bedFields.clear();
+ }
+}
+// END Intersect
- if ((bedLine.find("track") != string::npos) || (bedLine.find("browser") != string::npos)) {
- continue;
- }
- else {
- // split the current line into ditinct fields
- vector<string> bedFields;
- Tokenize(bedLine,bedFields);
- lineNum++;
-
- // find the overlaps with B if it's a valid BED entry.
- if (bedA->parseBedLine(a, bedFields, lineNum)) {
- vector<BED> hits;
- FindOverlaps(a, hits);
- }
- }
+void BedSubtract::DetermineBedInput() {
+ if (bedA->bedFile != "stdin") { // process a file
+ ifstream beds(bedA->bedFile.c_str(), ios::in);
+ if ( !beds ) {
+ cerr << "Error: The requested bed file (" << bedA->bedFile << ") could not be opened. Exiting!" << endl;
+ exit (1);
}
+ SubtractBed(beds);
+ }
+ else { // process stdin
+ SubtractBed(cin);
}
}
-// END Intersect
-
diff --git a/src/subtractBed/subtractBed.h b/src/subtractBed/subtractBed.h
index a7eb54ebccdfcbb4f9df5b11034de51fe9c92570..b11d86198798d993c861c2a1cacac2d802b8223e 100755
--- a/src/subtractBed/subtractBed.h
+++ b/src/subtractBed/subtractBed.h
@@ -1,3 +1,14 @@
+/*****************************************************************************
+ subtractBed.h
+
+ (c) 2009 - Aaron Quinlan
+ Hall Laboratory
+ Department of Biochemistry and Molecular Genetics
+ University of Virginia
+ aaronquinlan@gmail.com
+
+ Licenced under the GNU General Public License 2.0+ license.
+******************************************************************************/
#ifndef SUBTRACTBED_H
#define SUBTRACTBED_H
@@ -25,8 +36,8 @@ public:
void reportA(BED &);
void FindOverlaps(BED &, vector<BED> &);
- void SubtractBed();
-
+ void SubtractBed(istream &bedInput);
+ void DetermineBedInput();
private:
diff --git a/src/subtractBed/subtractMain.cpp b/src/subtractBed/subtractMain.cpp
index 6e95537b86c281bf11904bc278324d8077b249d1..592a78c8b158bf58c5c33d136eb7039b3a8a5d72 100755
--- a/src/subtractBed/subtractMain.cpp
+++ b/src/subtractBed/subtractMain.cpp
@@ -1,3 +1,14 @@
+/*****************************************************************************
+ subtractMain.cpp
+
+ (c) 2009 - Aaron Quinlan
+ Hall Laboratory
+ Department of Biochemistry and Molecular Genetics
+ University of Virginia
+ aaronquinlan@gmail.com
+
+ Licenced under the GNU General Public License 2.0+ license.
+******************************************************************************/
#include "subtractBed.h"
#include "version.h"
@@ -86,7 +97,7 @@ int main(int argc, char* argv[]) {
if (!showHelp) {
BedSubtract *bs = new BedSubtract(bedAFile, bedBFile, overlapFraction, forceStrand);
- bs->SubtractBed();
+ bs->DetermineBedInput();
return 0;
}
else {
@@ -101,17 +112,19 @@ void ShowHelp(void) {
cerr << " Aaron Quinlan, Ph.D. (aaronquinlan@gmail.com) " << endl ;
cerr << " Hall Laboratory, University of Virginia" << endl;
cerr << "===============================================" << endl << endl;
- cerr << "Description: Removes the portion(s) of an interval that is overlapped by another feature(s). " << endl << endl;
+ cerr << "SUMMARY: Removes the portion(s) of an interval that is overlapped" << endl;
+ cerr << "\t by another feature(s)." << endl << endl;
- cerr << "Usage: " << PROGRAM_NAME << " [OPTIONS] -a <a.bed> -b <b.bed>" << endl << endl;
+ cerr << "USAGE: " << PROGRAM_NAME << " [OPTIONS] -a <a.bed> -b <b.bed>" << endl << endl;
cerr << "OPTIONS: " << endl;
- cerr << "\t" << "-s\t\t\t" << "Force strandedness. Only report hits in B that overlap A on the same strand." << endl << "\t\t\t\tBy default, overlaps are reported without respect to strand." << endl << endl;
- cerr << "\t" << "-f (e.g. 0.05)\t\t" << "Minimum overlap req'd as fraction of a.bed." << endl << "\t\t\t\tDefault is 1E-9 (effectively 1bp)." << endl << endl;
+ cerr << " " << "-f\t" << "Minimum overlap required as a fraction of A." << endl;
+ cerr << "\t- Default is 1E-9 (i.e., 1bp)." << endl << endl;
+
+ cerr << " " << "-s\t" << "Force strandedness. That is, only report hits in B that" << endl;
+ cerr << "\toverlap A on the same strand." << endl;
+ cerr << "\t- By default, overlaps are reported without respect to strand." << endl << endl;
- cerr << "NOTES: " << endl;
- cerr << "\t" << "-i stdin\t\t" << "Allows BED file A to be read from stdin. E.g.: cat a.bed | subtractBed -a stdin -b B.bed" << endl << endl;
- cerr << "\t***Only tab-delimited BED3 - BED6 formats allowed.***"<< endl << endl;
// end the program here
exit(1);
diff --git a/src/windowBed/windowMain.cpp b/src/windowBed/windowMain.cpp
index 4e5a7502fba5321c00b5c0c6ff47cff866b70fc8..fd2641d2e5141e6a06b0d4a2bf4210ff6f1dfa64 100755
--- a/src/windowBed/windowMain.cpp
+++ b/src/windowBed/windowMain.cpp
@@ -168,44 +168,44 @@ void ShowHelp(void) {
cerr << " Hall Laboratory, University of Virginia" << endl;
cerr << "===============================================" << endl << endl;
cerr << "SUMMARY: Examines a \"window\" around each feature in A and" << endl;
- cerr << " reports all features in B that overlap the window. For each" << endl;
- cerr << " overlap the entire entry in A and B are reported." << endl << endl;
+ cerr << "\t reports all features in B that overlap the window. For each" << endl;
+ cerr << "\t overlap the entire entry in A and B are reported." << endl << endl;
- cerr << "USAGE: " << PROGRAM_NAME << " [OPTIONS] -a <a.bed> -b <b.bed>" << endl << endl;
+ cerr << "USAGE: " << PROGRAM_NAME << " [OPTIONS] -a <a.bed> -b <b.bed>" << endl << endl;
cerr << "OPTIONS: " << endl;
cerr << " " << "-w\t" << "Base pairs added upstream and downstream of each entry" << endl;
cerr << "\tin A when searching for overlaps in B." << endl;
- cerr << "\t - Creates symterical \"windows\" around A." << endl;
- cerr << "\t - Default is 1000 bp." << endl << endl;
+ cerr << "\t- Creates symterical \"windows\" around A." << endl;
+ cerr << "\t- Default is 1000 bp." << endl << endl;
cerr << " " << "-l\t" << "Base pairs added upstream (left of) of each entry" << endl;
cerr << "\tin A when searching for overlaps in B." << endl;
- cerr << "\t - Allows one to define assymterical \"windows\"." << endl;
- cerr << "\t - Default is 1000 bp." << endl << endl;
+ cerr << "\t- Allows one to define assymterical \"windows\"." << endl;
+ cerr << "\t- Default is 1000 bp." << endl << endl;
cerr << " " << "-r\t" << "Base pairs added downstream (right of) of each entry" << endl;
cerr << "\tin A when searching for overlaps in B." << endl;
- cerr << "\t - Allows one to define assymterical \"windows\"." << endl;
- cerr << "\t - Default is 1000 bp." << endl << endl;
+ cerr << "\t- Allows one to define assymterical \"windows\"." << endl;
+ cerr << "\t- Default is 1000 bp." << endl << endl;
cerr << " " << "-sw\t" << "Define -l and -r based on strand. For example if used, -l 500" << endl;
cerr << "\tfor a negative-stranded feature will add 500 bp downstream." << endl;
- cerr << "\t - Default = disabled." << endl << endl;
+ cerr << "\t- Default = disabled." << endl << endl;
cerr << " " << "-sm\t" << "Only report hits in B that overlap A on the same strand." << endl;
- cerr << "\t - By default, overlaps are reported without respect to strand." << endl << endl;
+ cerr << "\t- By default, overlaps are reported without respect to strand." << endl << endl;
cerr << " " << "-u\t" << "Write the original A entry _once_ if _any_ overlaps found in B." << endl;
- cerr << "\t - In other words, just report the fact >=1 hit was found." << endl << endl;
+ cerr << "\t- In other words, just report the fact >=1 hit was found." << endl << endl;
cerr << " " << "-c\t" << "For each entry in A, report the number of overlaps with B." << endl;
- cerr << "\t - Reports 0 for A entries that have no overlap with B." << endl;
- cerr << "\t - Overlaps restricted by -f." << endl << endl;
+ cerr << "\t- Reports 0 for A entries that have no overlap with B." << endl;
+ cerr << "\t- Overlaps restricted by -f." << endl << endl;
cerr << " " << "-v\t" << "Only report those entries in A that have _no overlaps_ with B." << endl;
- cerr << "\t - Similar to \"grep -v.\"" << endl << endl;
+ cerr << "\t- Similar to \"grep -v.\"" << endl << endl;
// end the program here
exit(1);