From 258cb98daa8f7d06d0e419b17d3b14e9fd577ea5 Mon Sep 17 00:00:00 2001
From: Vincent Danjean <Vincent.Danjean@ens-lyon.org>
Date: Sat, 13 Jan 2018 22:23:58 +0100
Subject: [PATCH] Disabling pipefail when the output is explicitely troncated

"grep -m 1" stop after the first found occurence.
"samtools view" stop with an error if its output is closed.
With pipefail, the samtools error is taken into account by the lumpyexpress script that immediately exits

The error do not always occurs:
if the information looked after by grep is at the end of the BAM file and samtools is able to complete before its output is closed, then no error occurs
But we have several BAM files that triggered this error.
---
 scripts/lumpyexpress | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/scripts/lumpyexpress b/scripts/lumpyexpress
index b8b7625..721307a 100755
--- a/scripts/lumpyexpress
+++ b/scripts/lumpyexpress
@@ -401,7 +401,9 @@ then
 
 	    # generate discordant pair string for LUMPY
 	    DISC_BAM=$TEMP_DIR/$OUTBASE.sample$(($i+1)).discordants.bam
+	    set +o pipefail
 	    DISC_SAMPLE=`$SAMT view $EXT_OPT -H $FULL_BAM | grep -m 1 "^@RG" | gawk -v i=$i '{ for (j=1;j<=NF;++j) {if ($j~"^SM:") { gsub("^SM:","",$j); print $j } } }'`
+	    set -o pipefail
 	    RG_STRING=`echo "${LIB_RG_LIST[$j]}" | sed 's/,/,read_group:/g' | sed 's/^/read_group:/g'`
 	    MEAN=`cat ${TEMP_DIR}/$OUTBASE.sample$(($i+1)).lib$(($j+1)).insert.stats | tr '\t' '\n' | grep "^mean" | sed 's/mean\://g'`
 	    STDEV=`cat ${TEMP_DIR}/$OUTBASE.sample$(($i+1)).lib$(($j+1)).insert.stats | tr '\t' '\n' | grep "^stdev" | sed 's/stdev\://g'`
@@ -409,7 +411,9 @@ then
 
 	    # generate split-read string for LUMPY
 	    SPL_BAM=$TEMP_DIR/$OUTBASE.sample$(($i+1)).splitters.bam
+	    set +o pipefail
 	    SPL_SAMPLE=`$SAMT view $EXT_OPT -H $FULL_BAM | grep -m 1 "^@RG" | gawk -v i=$i '{ for (j=1;j<=NF;++j) {if ($j~"^SM:") { gsub("^SM:","",$j); print $j } } }'`
+	    set -o pipefail
 	    LUMPY_SPL_STRING="$LUMPY_SPL_STRING -sr bam_file:${SPL_BAM},back_distance:10,min_mapping_threshold:20,weight:1,id:${SPL_SAMPLE},min_clip:20,${RG_STRING}"
 	done
 
@@ -487,7 +491,9 @@ else
 	echo "done"
 
 	# construct LUMPY_SPL_STRING
+	set +o pipefail
 	SPL_SAMPLE=`$SAMT view -H $SPL_BAM | grep -m 1 "^@RG" | gawk -v i=$i '{ for (j=1;j<=NF;++j) {if ($j~"^SM:") { gsub("^SM:","",$j); print $j } } }'`
+	set -o pipefail
 	LUMPY_SPL_STRING="$LUMPY_SPL_STRING -sr bam_file:${SPL_BAM},back_distance:10,min_mapping_threshold:20,weight:1,id:${SPL_SAMPLE},min_clip:20"
 
 	# construct LUMPY_DISC_STRING
@@ -495,7 +501,9 @@ else
 	do
 	    echo $(( ${#FULL_BAM_LIST[@]}-1 ))
 	    DISC_BAM=${DISC_BAM_LIST[$i]}
+	    set +o pipefail
 	    DISC_SAMPLE=`$SAMT view -H $DISC_BAM | grep -m 1 "^@RG" | gawk -v i=$i '{ for (j=1;j<=NF;++j) {if ($j~"^SM:") { gsub("^SM:","",$j); print $j } } }'`
+	    set -o pipefail
 	    MEAN=`cat ${TEMP_DIR}/$OUTBASE.sample$(($i+1)).lib$(($j+1)).insert.stats | tr '\t' '\n' | grep "^mean" | sed 's/mean\://g'`
 	    STDEV=`cat ${TEMP_DIR}/$OUTBASE.sample$(($i+1)).lib$(($j+1)).insert.stats | tr '\t' '\n' | grep "^stdev" | sed 's/stdev\://g'`
 	    RG_STRING=`echo "${LIB_RG_LIST[$j]}" | sed 's/,/,read_group:/g' | sed 's/^/read_group:/g'`