fix blacklist bug (lines matching chr[\dXY]_* were included)

kundajelab · Dec 20, 2016 · a7b053e · a7b053e
1 parent e154168
commit a7b053e
Show file tree

Hide file tree

Showing 4 changed files with 8 additions and 8 deletions.
diff --git a/modules/callpeak_bigbed.bds b/modules/callpeak_bigbed.bds
@@ -30,11 +30,11 @@ string peak_to_bigbed( string filetype, string peak, string o_dir, string group
 		sys $shcmd_init
 
 		sys zcat $peak | sort -k1,1 -k2,2n > $bigbed.tmp
-		// sys cat $chrsz | grep -P 'chr[0-9XY]+(?!_)' > $bigbed.chrsz.tmp
-		// sys bedToBigBed $bed_param $bigbed.tmp $bigbed.chrsz.tmp $bigbed
-		// sys rm -f $bigbed.tmp $bigbed.chrsz.tmp
-		sys bedToBigBed $bed_param $bigbed.tmp $chrsz $bigbed
-		sys rm -f $bigbed.tmp
+		sys cat $chrsz | grep -P 'chr[\dXY]+[ \t]' > $bigbed.chrsz.tmp
+		sys bedToBigBed $bed_param $bigbed.tmp $bigbed.chrsz.tmp $bigbed
+		sys rm -f $bigbed.tmp $bigbed.chrsz.tmp
+		// sys bedToBigBed $bed_param $bigbed.tmp $chrsz $bigbed
+		// sys rm -f $bigbed.tmp
 
 		sys $shcmd_finalize
 	}

diff --git a/modules/callpeak_blacklist_filter.bds b/modules/callpeak_blacklist_filter.bds
@@ -30,7 +30,7 @@ string blacklist_filter_peak( string filetype, string peak, string o_dir, string
 
 		sys bedtools intersect -v -a $peak -b $blacklist \
 			| awk 'BEGIN{OFS="\t"} {if ($5>1000) $5=1000; print $0}' \
-			| grep -P 'chr[0-9XY]+(?!_)' | gzip -nc > $filtered
+			| grep -P 'chr[\dXY]+[ \t]' | gzip -nc > $filtered
 
 		sys $shcmd_finalize
 	}

diff --git a/modules/callpeak_idr.bds b/modules/callpeak_idr.bds
@@ -51,7 +51,7 @@ string[] idr2( string peak1, string peak2, string pooled, string idr_thresh, str
 
 	if ( path_exists( blacklist ) ) {
 		cmd1 = "bedtools intersect -v -a $peak_idr_trk_tmp -b $blacklist"+\
-			" | grep -P 'chr[0-9XY]+(?!_)'"+\
+			" | grep -P 'chr[\dXY]+[ \t]'"+\
 			" | awk 'BEGIN{OFS=\"\t\"} {if ($5>1000) $5=1000; print $0}'"+\
 			" | gzip -nc > $filt_peak_idr_trk_tmp"
 		cmd2 = "zcat $filt_peak_idr_trk_tmp | awk 'BEGIN{OFS=\"\t\"} {print $1,$2,$3,$4,$5,$6,$7,$8,$9,$10}' | gzip -nc > $filt_peak_idr"

diff --git a/modules/callpeak_naive_overlap.bds b/modules/callpeak_naive_overlap.bds
@@ -40,7 +40,7 @@ string naive_overlap_peak( string filetype, string peak_pooled, string peak_rep1
 	string cmd1
 	if ( path_exists( blacklist ) ) {
 		cmd1 = "bedtools intersect -v -a $peak_final -b $blacklist "+\
-			" | grep -P 'chr[0-9XY]+(?!_)'"+\
+			" | grep -P 'chr[\dXY]+[ \t]'"+\
 			" | awk 'BEGIN{OFS=\"\t\"} {if ($5>1000) $5=1000; print $0}'"+\
 			" | gzip -nc > $filt_peak_final"
 	}