Merge pull request #24 from FredHutch/mpmeers-patch-191118

Update to SEACR v1.3

README.md

@@ -15,9 +15,14 @@ A web interface for SEACR analysis can be found at https://seacr.fredhutch.org
 
 ## Recent changes
 
+### v1.3
+
+- Fixed a bug in which the bedgraph line thresholding added in v1.2 was failing for some datasets.
+- Added a check to filter out any input bedgraph lines containing zero signal.
+
 ### v1.2
 
-- Fixed a bug in lines 166 and 168 in which misplaced brackets caused the misreporting of the max signal region terminal coordinate for merged signal blocks
+- Fixed a bug in lines 166 and 168 in which misplaced brackets caused the misreporting of the max signal region terminal coordinate for merged signal blocks.
 - Added a counter to keep track of the number of component bedgraph lines that compose each signal block, and a function to calculate the minimum threshold of lines per signal block at which there is a smaller percentage of target signal blocks remaining than control. This is meant to be used as a filter for signal blocks that pass the total signal threshold despite being composed of very few bedgraph lines, which are unlikely to be true peaks.
 - Changed how the dataframe for density plotting is truncated (previously a hard-coded 90% cutoff): a dataframe of list quantile (i.e. line #/max line#) vs. value quantile (i.e. value/max value) is derived, and the threshold is selected by finding the dataframe pair for which the orthogonal distance below the line defined by (0,0);(1,1) is maximized.
 
@@ -28,7 +33,7 @@ A web interface for SEACR analysis can be found at https://seacr.fredhutch.org
 
 ## Usage: 
 
-	bash SEACR_1.2.sh experimental bedgraph [control bedgraph | numeric threshold] ["norm" | "non"] ["relaxed" | "stringent"] output prefix
+	bash SEACR_1.3.sh experimental bedgraph [control bedgraph | numeric threshold] ["norm" | "non"] ["relaxed" | "stringent"] output prefix
 
 ## Description of input fields:
 
@@ -76,11 +81,11 @@ Field 6: Region representing the farthest upstream and farthest downstream bases
 
 ## Examples:
 
-	bash SEACR_1.2.sh target.bedgraph IgG.bedgraph norm stringent output
+	bash SEACR_1.3.sh target.bedgraph IgG.bedgraph norm stringent output
 Calls enriched regions in target data using normalized IgG control track with stringent threshold
 
-	bash SEACR_1.2.sh target.bedgraph IgG.bedgraph non relaxed output
+	bash SEACR_1.3.sh target.bedgraph IgG.bedgraph non relaxed output
 Calls enriched regions in target data using non-normalized IgG control track with relaxed threshold
 
-	bash SEACR_1.2.sh target.bedgraph 0.01 non stringent output
+	bash SEACR_1.3.sh target.bedgraph 0.01 non stringent output
 Calls enriched regions in target data by selecting the top 1% of regions by AUC

SEACR_1.2.R → SEACR_1.3.R

@@ -125,7 +125,11 @@ if(is.na(numtest)){ ## If 2nd field is a bedgraph, calculate empirical threshold
 	both2<-c(expmax,ctrlmax)
 	d<-sort(unique(both2))
 	pctremain2<-function(x) 1-(ecdf(expmax)(x)-ecdf(ctrlmax)(x))
-	d0<-min(d[pctremain2(d) > 1])
+	if(length(d[pctremain2(d) > 1]) > 0){
+		d0<-min(d[pctremain2(d) > 1])
+	}else{
+		d0<-1
+	}
 	invis <- gc(verbose=FALSE)
 	fdr<-c(1-pctremain(x0[1]), 1-pctremain(z0[1])) ## New for SEACR_1.1
 }else{ ## If 2nd field is numeric, calculate percentile threshold

SEACR_1.2.sh → SEACR_1.3.sh

@@ -7,7 +7,7 @@ then
 	echo "
 	SEACR: Sparse Enrichment Analysis for CUT&RUN
 	
-	Usage: bash SEACR_1.2.sh <experimental bedgraph>.bg [<control bedgraph>.bg | <FDR threshold>] ["norm" | "non"] ["relaxed" | "stringent"] output prefix
+	Usage: bash SEACR_1.3.sh <experimental bedgraph>.bg [<control bedgraph>.bg | <FDR threshold>] ["norm" | "non"] ["relaxed" | "stringent"] output prefix
 	
 	Description of input fields:
 	
@@ -42,12 +42,12 @@ then
 	Field 6: Region representing the farthest upstream and farthest downstream bases within the denoted coordinates that are represented by the maximum bedgraph signal
 	
 	Examples:
-	bash SEACR_1.2.sh target.bedgraph IgG.bedgraph norm stringent output
+	bash SEACR_1.3.sh target.bedgraph IgG.bedgraph norm stringent output
 	Calls enriched regions in target data using normalized IgG control track with stringent threshold
 	
-	bash SEACR_1.2.sh target.bedgraph IgG.bedgraph non relaxed output
+	bash SEACR_1.3.sh target.bedgraph IgG.bedgraph non relaxed output
 	Calls enriched regions in target data using non-normalized IgG control track with relaxed threshold
-	bash SEACR_1.2.sh target.bedgraph 0.01 non stringent output
+	bash SEACR_1.3.sh target.bedgraph 0.01 non stringent output
 	Calls enriched regions in target data by selecting the top 1% of regions by area under the curve (AUC)
 	"
 	exit 1
@@ -98,16 +98,16 @@ fi
 
 echo "Creating experimental AUC file: $(date)"
 
-awk 'BEGIN{s=1}; {if(s==1){s++}else if(s==2){chr=$1; start=$2; stop=$3; max=$4; coord=$1":"$2"-"$3; auc=$4*($3-$2); num=1; s++}else{if(chr==$1 && $2==stop){num++; stop=$3; auc=auc+($4*($3-$2)); if ($4 > max){max=$4; coord=$1":"$2"-"$3
-}else if($4 == max){split(coord,t,"-"); coord=t[1]"-"$3}}else{print chr"\t"start"\t"stop"\t"auc"\t"max"\t"coord"\t"num; chr=$1; start=$2; stop=$3; max=$4; coord=$1":"$2"-"$3; auc=$4*($3-$2); num=1}}}' $1 > $password.auc.bed
+awk 'BEGIN{s=1}; {if(s==1){s++}else if(s==2){if($4 > 0){chr=$1; start=$2; stop=$3; max=$4; coord=$1":"$2"-"$3; auc=$4*($3-$2); num=1; s++}}else{if($4 > 0){if(chr==$1 && $2==stop){num++; stop=$3; auc=auc+($4*($3-$2)); if ($4 > max){max=$4; coord=$1":"$2"-"$3
+}else if($4 == max){split(coord,t,"-"); coord=t[1]"-"$3}}else{print chr"\t"start"\t"stop"\t"auc"\t"max"\t"coord"\t"num; chr=$1; start=$2; stop=$3; max=$4; coord=$1":"$2"-"$3; auc=$4*($3-$2); num=1}}}}' $1 > $password.auc.bed
 cut -f 4,7 $password.auc.bed > $password.auc
 
 if [[ -f $2 ]]
 then
   echo "Creating control AUC file: $(date)"
 
-  awk 'BEGIN{s=1}; {if(s==1){s++}else if(s==2){chr=$1; start=$2; stop=$3; max=$4; coord=$1":"$2"-"$3; auc=$4*($3-$2); num=1; s++}else{if(chr==$1 && $2==stop){num++; stop=$3; auc=auc+($4*($3-$2)); if ($4 > max){max=$4; coord=$1":"$2"-"
-$3}else if($4 == max){split(coord,t,"-"); coord=t[1]"-"$3}}else{print chr"\t"start"\t"stop"\t"auc"\t"max"\t"coord"\t"num; chr=$1; start=$2; stop=$3; max=$4; coord=$1":"$2"-"$3; auc=$4*($3-$2); num=1}}}' $2 > $password2.auc.bed
+  awk 'BEGIN{s=1}; {if(s==1){s++}else if(s==2){if($4 > 0){chr=$1; start=$2; stop=$3; max=$4; coord=$1":"$2"-"$3; auc=$4*($3-$2); num=1; s++}}else{if($4 > 0){if(chr==$1 && $2==stop){num++; stop=$3; auc=auc+($4*($3-$2)); if ($4 > max){max=$4; coord=$1":"$2"-"
+$3}else if($4 == max){split(coord,t,"-"); coord=t[1]"-"$3}}else{print chr"\t"start"\t"stop"\t"auc"\t"max"\t"coord"\t"num; chr=$1; start=$2; stop=$3; max=$4; coord=$1":"$2"-"$3; auc=$4*($3-$2); num=1}}}}' $2 > $password2.auc.bed
   cut -f 4,7 $password2.auc.bed > $password2.auc
 fi
 
@@ -119,14 +119,14 @@ path=`dirname $0`
 if [[ -f $2 ]] && [[ $norm == "norm" ]]
 then
 	echo "Calculating threshold using normalized control: $(date)"
-	Rscript $path/SEACR_1.2.R --exp=$password.auc --ctrl=$password2.auc --norm=yes --output=$password
+	Rscript $path/SEACR_1.3.R --exp=$password.auc --ctrl=$password2.auc --norm=yes --output=$password
 elif [[ -f $2 ]]
 then
 	echo "Calculating threshold using non-normalized control: $(date)"
-	Rscript $path/SEACR_1.2.R --exp=$password.auc --ctrl=$password2.auc --norm=no --output=$password
+	Rscript $path/SEACR_1.3.R --exp=$password.auc --ctrl=$password2.auc --norm=no --output=$password
 else
 	echo "Using user-provided threshold: $(date)"
-	Rscript $path/SEACR_1.2.R --exp=$password.auc --ctrl=$2 --norm=no --output=$password
+	Rscript $path/SEACR_1.3.R --exp=$password.auc --ctrl=$2 --norm=no --output=$password
 fi
 
 fdr=`cat $password.fdr.txt | sed -n '1p'`			## Added 5/15/19 for SEACR_1.1