From 96e2ac601c1392b6d62d1ff2c0978e6860ff7f2d Mon Sep 17 00:00:00 2001
From: Aaron <aaronquinlan@gmail.com>
Date: Thu, 3 Mar 2011 10:57:34 -0500
Subject: [PATCH] Optionally set masking character (-mc) in maskFastaFromBed.
Thanks to Can Alkan.
---
src/maskFastaFromBed/maskFastaFromBed.cpp | 24 ++++++++----------
src/maskFastaFromBed/maskFastaFromBed.h | 4 ++-
src/maskFastaFromBed/maskFastaFromBedMain.cpp | 25 +++++++++++++++----
3 files changed, 33 insertions(+), 20 deletions(-)
diff --git a/src/maskFastaFromBed/maskFastaFromBed.cpp b/src/maskFastaFromBed/maskFastaFromBed.cpp
index 654ff6db..2bab1226 100644
--- a/src/maskFastaFromBed/maskFastaFromBed.cpp
+++ b/src/maskFastaFromBed/maskFastaFromBed.cpp
@@ -13,21 +13,17 @@
#include "maskFastaFromBed.h"
-MaskFastaFromBed::MaskFastaFromBed(string &fastaInFile, string &bedFile, string &fastaOutFile, bool &softMask) {
-
- _softMask = false;
- if (softMask) {
- _softMask = true;
- }
-
- _fastaInFile = fastaInFile;
- _bedFile = bedFile;
+MaskFastaFromBed::MaskFastaFromBed(const string &fastaInFile, const string &bedFile,
+ const string &fastaOutFile, bool softMask, char maskChar) {
+ _softMask = softMask;
+ _fastaInFile = fastaInFile;
+ _bedFile = bedFile;
_fastaOutFile = fastaOutFile;
-
- _bed = new BedFile(_bedFile);
+ _maskChar = maskChar;
+ _bed = new BedFile(_bedFile);
_bed->loadBedFileIntoMapNoBin();
-
+ // start masking.
MaskFasta();
}
@@ -102,7 +98,7 @@ void MaskFastaFromBed::MaskFasta() {
currDNA.replace(start, length, replacement);
}
else {
- string hardmask(length, 'N');
+ string hardmask(length, _maskChar);
currDNA.replace(start, length, hardmask);
}
}
@@ -133,7 +129,7 @@ void MaskFastaFromBed::MaskFasta() {
currDNA.replace(start, length, replacement);
}
else {
- string hardmask(length, 'N');
+ string hardmask(length, _maskChar);
currDNA.replace(start, length, hardmask);
}
}
diff --git a/src/maskFastaFromBed/maskFastaFromBed.h b/src/maskFastaFromBed/maskFastaFromBed.h
index 1b7a3590..4cf68c84 100644
--- a/src/maskFastaFromBed/maskFastaFromBed.h
+++ b/src/maskFastaFromBed/maskFastaFromBed.h
@@ -28,7 +28,8 @@ class MaskFastaFromBed {
public:
// constructor
- MaskFastaFromBed(string &fastaInFile, string &bedFile, string &fastaOutFile, bool &softMask);
+ MaskFastaFromBed(const string &fastaInFile, const string &bedFile,
+ const string &fastaOutFile, bool softMask, char maskChar);
// destructor
~MaskFastaFromBed(void);
@@ -41,6 +42,7 @@ private:
string _fastaInFile;
string _bedFile;
string _fastaOutFile;
+ char _maskChar; // typically "N", but user's can choose something else, e.g., "X"
// instance of a bed file class.
BedFile *_bed;
diff --git a/src/maskFastaFromBed/maskFastaFromBedMain.cpp b/src/maskFastaFromBed/maskFastaFromBedMain.cpp
index 0978436d..7fce56bb 100644
--- a/src/maskFastaFromBed/maskFastaFromBedMain.cpp
+++ b/src/maskFastaFromBed/maskFastaFromBedMain.cpp
@@ -36,11 +36,12 @@ int main(int argc, char* argv[]) {
// output files
string fastaOutFile;
- // checks for existence of parameters
- bool haveFastaIn = false;
- bool haveBed = false;
+ // defaults for parameters
+ bool haveFastaIn = false;
+ bool haveBed = false;
bool haveFastaOut = false;
- bool softMask = false;
+ bool softMask = false;
+ char maskChar = 'N';
// check to see if we should print out some help
if(argc <= 1) showHelp = true;
@@ -85,6 +86,19 @@ int main(int argc, char* argv[]) {
else if(PARAMETER_CHECK("-soft", 5, parameterLength)) {
softMask = true;
}
+ else if(PARAMETER_CHECK("-mc", 3, parameterLength)) {
+ if ((i+1) < argc) {
+ string mask = argv[i + 1];
+ if (mask.size() > 1) {
+ cerr << "*****ERROR: The mask character (-mc) should be a single character.*****" << endl << endl;
+ showHelp = true;
+ }
+ else {
+ maskChar = mask[0];
+ }
+ i++;
+ }
+ }
else {
cerr << "*****ERROR: Unrecognized parameter: " << argv[i] << " *****" << endl << endl;
showHelp = true;
@@ -97,7 +111,7 @@ int main(int argc, char* argv[]) {
if (!showHelp) {
- MaskFastaFromBed *maskFasta = new MaskFastaFromBed(fastaInFile, bedFile, fastaOutFile, softMask);
+ MaskFastaFromBed *maskFasta = new MaskFastaFromBed(fastaInFile, bedFile, fastaOutFile, softMask, maskChar);
delete maskFasta;
return 0;
}
@@ -124,6 +138,7 @@ void ShowHelp(void) {
cerr << "\t-fo\tOutput FASTA file" << endl;
cerr << "\t-soft\tEnforce \"soft\" masking. That is, instead of masking with Ns," << endl;
cerr << "\t\tmask with lower-case bases." << endl;
+ cerr << "\t-mc\tReplace masking character. That is, instead of masking with Ns, use another character." << endl;
// end the program here
exit(1);
--
GitLab