HAP-280 more + fixed test cases

Author: Peter Krusche

configure.sh

@@ -41,12 +41,24 @@ else
     CONFIGTYPE="$2"
 fi
 
+PLATFORM='unknown'
+UNAMESTR=`uname`
+if [[ "$UNAMESTR" == 'Linux' ]]; then
+   PLATFORM='linux'
+elif [[ "$UNAMESTR" == 'FreeBSD' ]]; then
+   PLATFORM='freebsd'
+fi
+
 if [[ -z $CONFIGTYPE ]]; then
-	if [[ "$(hostname)" == *chuk.illumina.com ]] || [[ "$(hostname)" == ukch-* ]] || [[ "$(hostname)" == ussd-prd-lndt* ]]; then
-	    echo "using Illumina configuration."
-	    SPECIALCONFIG="${DIR}/src/sh/illumina-setup.sh"
-	    . ${DIR}/src/sh/illumina-setup.sh
-	fi
+    if [[ "$PLATFORM" == "linux" ]] && [[ -d /illumina ]]; then
+        if [[ "$(hostname)" == *.illumina.com ]] || \
+           [[ "$(hostname)" == ukch-* ]] || \
+           [[ "$(hostname)" == ussd-prd-lndt* ]]; then
+            echo "using Illumina configuration."
+            SPECIALCONFIG="${DIR}/src/sh/illumina-setup.sh"
+            . ${DIR}/src/sh/illumina-setup.sh
+        fi        
+    fi
 else
     echo "using $CONFIGTYPE configuration."
     SPECIALCONFIG="${DIR}/src/sh/$CONFIGTYPE-setup.sh"

example/callsonly/call_merge.vcf

@@ -8,8 +8,8 @@
 ##contig=<ID=21,length=48129895>
 #CHROM	POS	ID	REF	ALT	QUAL	FILTER	INFO	FORMAT	NA12877	NA12878	NA12882
 chr21	20000000	.	N	.	.	PASS	END=20000100	GT:DP:GQ	./.:.:.	./.:.:.	0/0:20:10
-chr21	20000020	.	N	.	.	PASS	END=20000050	GT:DP:GQ	./.:.:.	0/0:30:20	./.:.:.
 chr21	20000020	.	A	T	.	PASS	.	GT:DP:GQ:AD	0/1:10:15:4,6	./.:.:.:.	./.:.:.:.
+chr21	20000021	.	N	.	.	PASS	END=20000050	GT:DP:GQ	./.:.:.	0/0:30:20	./.:.:.
 chr21	20000048	.	ATGG	A	.	PASS	.	GT:DP:GQ:AD	0/1:10:15:4,6	./.:.:.:.	./.:.:.:.
 chr21	20000051	.	N	.	.	PASS	END=20000060	GT:DP:GQ	./.:.:.	0/0:40:20	./.:.:.
 chr21	20000098	.	N	.	.	PASS	END=20000110	GT:DP:GQ	./.:.:.	0/0:70:21	./.:.:.

example/callsonly/call_merge.vcf.gz

@@ -0,0 +1,5 @@
+Type,Filter,TRUTH.TOTAL,TRUTH.TP,TRUTH.FN,QUERY.TOTAL,QUERY.FP,QUERY.UNK,FP.gt,METRIC.Recall,METRIC.Precision,METRIC.Frac_NA,TRUTH.TOTAL.TiTv_ratio,QUERY.TOTAL.TiTv_ratio,TRUTH.TOTAL.het_hom_ratio,QUERY.TOTAL.het_hom_ratio
+INDEL,ALL,9126,8532,594,14475,367,5375,18,0.934911,0.95967,0.37133,,,1.4244024458,1.93149834437
+INDEL,PASS,9126,8532,594,14449,366,5350,18,0.934911,0.959776,0.370268,,,1.4244024458,1.9267232457
+SNP,ALL,51016,50157,859,70641,130,20405,60,0.983162,0.997412,0.288855,2.08128510176,1.85179356817,1.59334960342,2.28952886901
+SNP,PASS,51016,50156,860,70551,130,20316,60,0.983143,0.997412,0.287962,2.08128510176,1.85242059323,1.59334960342,2.28640686183

example/callsonly/call_merge.vcf.gz.tbi

@@ -68,12 +68,12 @@ def main():
         print >> sys.stderr, "-------------------------"
         sys.exit(1)
 
+    print "Comparing %i labels and %i metrics..." % (len(all_labels_1), len(all_keys_1))
+
     different_metrics = []
 
     for label in all_labels_1:
         for metric in all_keys_1:
-            print label + "/" + metric
-
             v1 = data1[label][metric]
             v2 = data2[label][metric]
 

example/happy/expected-stratified.extended.csv

@@ -38,7 +38,7 @@ ${HCDIR}/multimerge ${DIR}/../../example/homref/homref.vcf.gz \
 diff -I ^# ${TF} ${DIR}/../../example/homref/expected_merge.vcf
 
 if [ $? -ne 0 ]; then
-	echo "GVCF homref test FAILED. You can inspect ${TF} for the failed result."
+	echo "GVCF homref test FAILED. You can inspect ${TF} for the failed result. //  diff -I ^# ${TF} ${DIR}/../../example/homref/expected_merge.vcf"
 	exit 1
 else
 	echo "GVCF homref test SUCCEEDED."
@@ -67,7 +67,7 @@ ${HCDIR}/multimerge ${DIR}/../../example/callsonly/call_merge.vcf.gz:* \
 diff ${TF} ${DIR}/../../example/callsonly/expected_callsonly.vcf
 
 if [ $? -ne 0 ]; then
-	echo "GVCF homref + Variants test FAILED. You can inspect ${TF} for the failed result."
+	echo "GVCF homref + Variants test FAILED. You can inspect ${TF} for the failed result. // diff ${TF} ${DIR}/../../example/callsonly/expected_callsonly.vcf"
 	exit 1
 else
 	echo "GVCF homref + Variants test SUCCEEDED."

example/happy/expected-stratified.summary.csv

@@ -0,0 +1,49 @@
+#!/bin/bash
+
+# Quantification and GA4GH intermediate file format compliance
+#
+
+set +e
+
+DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" && pwd )"
+. ${DIR}/detect_vars.sh
+
+echo "Quantify stratified counting test for ${HCVERSION} from ${HCDIR}"
+
+HG38=${DIR}/../../example/happy/hg38.chr21.fa
+
+TMP_OUT=`mktemp -t happy.XXXXXXXXXX`
+
+# run hap.py on NA12878 test file
+# note this only works if we don't change the confident regions
+# based on the truthset
+${PYTHON} ${HCDIR}/hap.py \
+			 	-l chr21 \
+			 	${DIR}/../../example/happy/PG_NA12878_hg38-chr21.vcf.gz \
+			 	${DIR}/../../example/happy/NA12878-GATK3-chr21.vcf.gz \
+			 	-f ${DIR}/../../example/happy/PG_Conf_hg38-chr21.bed.gz \
+			 	-r ${HG38} \
+			 	-o ${TMP_OUT} \
+                --stratification ${DIR}/../../example/happy/stratification.tsv \
+			 	--force-interactive
+
+if [[ $? != 0 ]]; then
+	echo "hap.py failed!"
+	exit 1
+fi
+
+# This script checks if the summary precision / recall figures have changed significantly
+${PYTHON} ${DIR}/compare_summaries.py ${TMP_OUT}.summary.csv ${DIR}/../../example/happy/expected-stratified.summary.csv
+if [[ $? != 0 ]]; then
+	echo "All summary differs! -- diff ${TMP_OUT}.summary.csv ${DIR}/../../example/happy/expected-stratified.summary.csv"
+	exit 1
+fi
+
+# This script checks if the extended precision / recall figures have changed significantly
+${PYTHON} ${DIR}/compare_extended.py ${TMP_OUT}.extended.csv ${DIR}/../../example/happy/expected-stratified.extended.csv
+if [[ $? != 0 ]]; then
+	echo "All extended differs! -- diff ${TMP_OUT}.extended.csv ${DIR}/../../example/happy/expected-stratified.extended.csv"
+	exit 1
+fi
+
+rm -rf ${TMP_OUT}.*

src/sh/compare_extended.py

@@ -215,6 +215,20 @@ else
 	echo "Quantify integration test SUCCEEDED!"
 fi
 
+##############################################################
+# Test GA4GH stratified quantification
+##############################################################
+
+/bin/bash ${DIR}/run_quantify_stratification_test.sh
+
+if [[ $? -ne 0 ]]; then
+	echo "Quantify stratification test FAILED!"
+	exit 1
+else
+	echo "Quantify stratification test SUCCEEDED!"
+fi
+
+
 ##############################################################
 # Test PG Counting
 ##############################################################