Merge pull request #92 from genenetwork/gemma-master

Fixed for a number of outstanding issues.

3 files changed, 78 insertions, 17 deletions

diff --git a/test/dev_test_suite.sh b/test/dev_test_suite.sh
index 9e49251..0fc4423 100755
--- a/test/dev_test_suite.sh
+++ b/test/dev_test_suite.sh
@@ -2,7 +2,6 @@
 
 gemma=../bin/gemma
 
-
 # Related to https://github.com/genetics-statistics/GEMMA/issues/78
 testBXDStandardRelatednessMatrixKSingularError() {
     outn=BXDerr
@@ -32,7 +31,8 @@ testBXDStandardRelatednessMatrixK() {
     assertEquals "-116.11" `perl -nle 'foreach $x (split(/\s+/,$_)) { $sum += sprintf("%.2f",(substr($x,,0,6))) } END { printf "%.2f",$sum }' $outfn`
 }
 
-testBXDMultivariateLinearMixedModel() {
+testBXDLMMLikelihoodRatio() {
+    outn=BXD_LMM_LR
     $gemma -g ../example/BXD_geno.txt.gz \
            -p ../example/BXD_pheno.txt \
            -c ../example/BXD_covariates2.txt \
@@ -40,12 +40,12 @@ testBXDMultivariateLinearMixedModel() {
            -k ./output/BXD.cXX.txt \
            -lmm 2 -maf 0.1 \
            -debug \
-           -o BXD_mvlmm
+           -o $outn
     assertEquals 0 $?
 
-    outfn=output/BXD_mvlmm.assoc.txt
-    assertEquals "65862" `wc -w < $outfn`
-    assertEquals "3088489421.94" `perl -nle 'foreach $x (split(/\s+/,$_)) { $sum += sprintf("%.2f",(substr($x,,0,6))) } END { printf "%.2f",$sum }' $outfn`
+    outfn=output/$outn.assoc.txt
+    assertEquals "80498" `wc -w < $outfn`
+    assertEquals "3088458212.93" `perl -nle 'foreach $x (split(/\s+/,$_)) { $sum += sprintf("%.2f",(substr($x,,0,6))) } END { printf "%.2f",$sum }' $outfn`
 }
 
 testCenteredRelatednessMatrixKLOCO1() {
@@ -65,10 +65,12 @@ testCenteredRelatednessMatrixKLOCO1() {
 testUnivariateLinearMixedModelLOCO1() {
     outn=mouse_hs1940_CD8_LOCO1_lmm
     rm -f output/$outn.*
-    $gemma -g ../example/mouse_hs1940.geno.txt.gz -p ../example/mouse_hs1940.pheno.txt \
+    $gemma -g ../example/mouse_hs1940.geno.txt.gz \
+           -p ../example/mouse_hs1940.pheno.txt \
 	   -n 1 \
 	   -loco 1 \
-           -a ../example/mouse_hs1940.anno.txt -k ./output/mouse_hs1940_LOCO1.cXX.txt \
+           -a ../example/mouse_hs1940.anno.txt \
+           -k ./output/mouse_hs1940_LOCO1.cXX.txt \
 	   -snps ../example/mouse_hs1940_snps.txt -lmm \
 	   -nind 400 \
 	   -debug \
@@ -78,7 +80,7 @@ testUnivariateLinearMixedModelLOCO1() {
     assertEquals 0 $?
     outfn=output/$outn.assoc.txt
     assertEquals "68" `wc -l < $outfn`
-    assertEquals "15465553.30" `perl -nle 'foreach $x (split(/\s+/,$_)) { $sum += sprintf("%.2f",(substr($x,,0,6))) } END { printf "%.2f",$sum }' $outfn`
+    assertEquals "15465346.22" `perl -nle 'foreach $x (split(/\s+/,$_)) { $sum += sprintf("%.2f",(substr($x,,0,6))) } END { printf "%.2f",$sum }' $outfn`
 }
 
 shunit2=`which shunit2`
diff --git a/test/lengthy_test_suite.sh b/test/lengthy_test_suite.sh
new file mode 100755
index 0000000..231475e
--- /dev/null
+++ b/test/lengthy_test_suite.sh
@@ -0,0 +1,55 @@
+#!/usr/bin/env bash
+#
+# Long running tests go here
+
+gemma=../bin/gemma
+
+testPlinkStandardRelatednessMatrixK() {
+    testname=testPlinkStandardRelatednessMatrixK
+    datadir=../example
+    outfn=output/$testname.sXX.txt
+    rm -f $outfn
+    $gemma -bfile $datadir/HLC \
+           -gk 2 -o $testname \
+           -debug
+    assertEquals 0 $?
+    assertEquals "427" `wc -l < $outfn`
+    assertEquals "-358.07" `perl -nle 'foreach $x (split(/\s+/,$_)) { $sum += sprintf("%.2f",(substr($x,,0,6))) } END { printf "%.2f",$sum }' $outfn`
+}
+
+testPlinkMultivariateLinearMixedModelMultiplePhenotypes_Issue58() {
+    echo "Long running test!"
+    # This test passes, but takes over 30 minutes to run!
+    # n=2 is original pheno in fam file
+    # n=1 is causal1
+    # n=3..12 is causal2
+    # n=13..22 is causal3
+    # -n 1 2 3 15 is independent
+    testname=testPlinkMultivariateLinearMixedModelMultiplePhenotypes
+    datadir=../example
+    $gemma -bfile $datadir/HLC \
+           -p $datadir/HLC.simu.pheno.txt \
+           -k output/testPlinkStandardRelatednessMatrixK.sXX.txt \
+           -lmm 1 \
+           -maf 0.1 \
+           -n 1 2 3 15 \
+           -c $datadir/HLC_covariates.txt \
+           -debug \
+           -o $testname
+    assertEquals 0 $?
+    outfn=output/$testname.assoc.txt
+    assertEquals "223243" `wc -l < $outfn`
+    assertEquals "89754977983.69" `perl -nle 'foreach $x (split(/\s+/,$_)) { $sum += sprintf("%.2f",(substr($x,,0,6))) } END { printf "%.2f",$sum }' $outfn`
+}
+
+shunit2=`which shunit2`
+
+if [ -x "$shunit2" ]; then
+    echo run system shunit2
+    . $shunit2
+elif [ -e ../contrib/shunit2-2.0.3/src/shell/shunit2 ]; then
+    echo run shunit2 provided in gemma repo
+    . ../contrib/shunit2-2.0.3/src/shell/shunit2
+else
+    echo "Can not find shunit2 - see INSTALL.md"
+fi
diff --git a/test/test_suite.sh b/test/test_suite.sh
index 44eb14c..350fc27 100755
--- a/test/test_suite.sh
+++ b/test/test_suite.sh
@@ -30,13 +30,13 @@ testUnivariateLinearMixedModelFullLOCO1() {
     assertEquals 0 $?
     outfn=output/$outn.assoc.txt
     assertEquals "951" `wc -l < $outfn`
-    assertEquals "267509369.79" `perl -nle 'foreach $x (split(/\s+/,$_)) { $sum += sprintf("%.2f",(substr($x,,0,6))) } END { printf "%.2f",$sum }' $outfn`
+    assertEquals "267507851.98" `perl -nle 'foreach $x (split(/\s+/,$_)) { $sum += sprintf("%.2f",(substr($x,,0,6))) } END { printf "%.2f",$sum }' $outfn`
 }
 
 testCenteredRelatednessMatrixK() {
     $gemma -g ../example/mouse_hs1940.geno.txt.gz \
            -p ../example/mouse_hs1940.pheno.txt \
-           -gk -o mouse_hs1940
+           -gk -o mouse_hs1940 -debug
     assertEquals 0 $?
     outfn=output/mouse_hs1940.cXX.txt
     assertEquals "1940" `wc -l < $outfn`
@@ -52,13 +52,14 @@ testUnivariateLinearMixedModel() {
            -a ../example/mouse_hs1940.anno.txt \
            -k ./output/mouse_hs1940.cXX.txt \
            -lmm \
-           -o mouse_hs1940_CD8_lmm
+           -o mouse_hs1940_CD8_lmm \
+           -debug
     assertEquals 0 $?
     grep "total computation time" < output/mouse_hs1940_CD8_lmm.log.txt
     assertEquals 0 $?
     outfn=output/mouse_hs1940_CD8_lmm.assoc.txt
-    assertEquals "118459" `wc -w < $outfn`
-    assertEquals "4038557453.62" `perl -nle 'foreach $x (split(/\s+/,$_)) { $sum += sprintf("%.2f",(substr($x,,0,6))) } END { printf "%.2f",$sum }' $outfn`
+    assertEquals "129228" `wc -w < $outfn`
+    assertEquals "4038540440.86" `perl -nle 'foreach $x (split(/\s+/,$_)) { $sum += sprintf("%.2f",(substr($x,,0,6))) } END { printf "%.2f",$sum }' $outfn`
 }
 
 testMultivariateLinearMixedModel() {
@@ -67,7 +68,8 @@ testMultivariateLinearMixedModel() {
            -n 1 6 \
            -a ../example/mouse_hs1940.anno.txt \
            -k ./output/mouse_hs1940.cXX.txt \
-           -lmm -o mouse_hs1940_CD8MCH_lmm
+           -lmm -o mouse_hs1940_CD8MCH_lmm \
+           -debug
     assertEquals 0 $?
 
     outfn=output/mouse_hs1940_CD8MCH_lmm.assoc.txt
@@ -81,7 +83,8 @@ testPlinkStandardRelatednessMatrixK() {
     outfn=output/$testname.sXX.txt
     rm -f $outfn
     $gemma -bfile $datadir/HLC \
-           -gk 2 -o $testname
+           -gk 2 -o $testname \
+           -debug
     assertEquals 0 $?
     assertEquals "427" `wc -l < $outfn`
     assertEquals "-358.07" `perl -nle 'foreach $x (split(/\s+/,$_)) { $sum += sprintf("%.2f",(substr($x,,0,6))) } END { printf "%.2f",$sum }' $outfn`
@@ -97,11 +100,12 @@ testPlinkMultivariateLinearMixedModel() {
            -lmm 1 \
            -maf 0.1 \
            -c $datadir/HLC_covariates.txt \
+           -debug \
            -o $testname
     assertEquals 0 $?
     outfn=output/$testname.assoc.txt
     assertEquals "223243" `wc -l < $outfn`
-    assertEquals "89756559859.06" `perl -nle 'foreach $x (split(/\s+/,$_)) { $sum += sprintf("%.2f",(substr($x,,0,6))) } END { printf "%.2f",$sum }' $outfn`
+    assertEquals "89757159113.77" `perl -nle 'foreach $x (split(/\s+/,$_)) { $sum += sprintf("%.2f",(substr($x,,0,6))) } END { printf "%.2f",$sum }' $outfn`
 }
 
 shunit2=`which shunit2`