zzh
/
gatk-3.8
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Now populating the refseq and dbsnp in awk instead of retrieving from firehose. 

Added refseq table to the pipeline object.


git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@5020 348d0f76-0448-11de-a6fe-93d51630548a
This commit is contained in:
kshakir 2011-01-19 18:19:10 +00:00
parent 55eb0387ac
commit c901fb6d70
2 changed files with 30 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
java/src/org/broadinstitute/sting/datasources/pipeline/PipelineProject.java
View File

@ -36,6 +36,7 @@ public class PipelineProject {
private File referenceFile; private File referenceFile;
private File intervalList; private File intervalList;
private File dbsnpFile; private File dbsnpFile;
private File refseqTable;
private Map<String, String> tags = new TreeMap<String, String>(); private Map<String, String> tags = new TreeMap<String, String>();
public String getName() { public String getName() {
@ -70,6 +71,14 @@ public class PipelineProject {
this.dbsnpFile = dbsnpFile; this.dbsnpFile = dbsnpFile;
} }
public File getRefseqTable() {
return refseqTable;
}
public void setRefseqTable(File refseqTable) {
this.refseqTable = refseqTable;
}
public Map<String, String> getTags() { public Map<String, String> getTags() {
return tags; return tags;
} }

26
shell/getFirehosePipelineYaml.sh
View File

@ -27,7 +27,7 @@ PIPELINE_YAML_FILE=$ENTITY_SET_ID.yaml
# Annotations to pull down from Firehose # Annotations to pull down from Firehose
FIREHOSE_ANNOTATIONS=(reference_file dbsnp_file interval_list \ FIREHOSE_ANNOTATIONS=(reference_file interval_list \
sample_id recalibrated_bam_file squid_project collaborator_id) sample_id recalibrated_bam_file squid_project collaborator_id)
# YAML templates # YAML templates
@ -37,8 +37,9 @@ PROJECT_YAML_TEMPLATE='"\n\
name: '"$ENTITY_SET_ID"',\n\ name: '"$ENTITY_SET_ID"',\n\
referenceFile: %s,\n\ referenceFile: %s,\n\
dbsnpFile: %s,\n\ dbsnpFile: %s,\n\
refseqTable: %s,\n\
intervalList: %s\n\ intervalList: %s\n\
},", $1, $2, $3' },", $1, dbsnp, refseq, $2'
SAMPLE_YAML_TEMPLATE='"\n\ SAMPLE_YAML_TEMPLATE='"\n\
{\n\ {\n\
@ -48,7 +49,7 @@ SAMPLE_YAML_TEMPLATE='"\n\
SQUIDProject: %s,\n\ SQUIDProject: %s,\n\
CollaboratorID: %s\n\ CollaboratorID: %s\n\
}\n\ }\n\
}", $4, $5, $6, $7' }", $3, $4, $5, $6'
TEST_AWK_COUNT=`echo '\n' | awk '{print $0}' | wc -c` TEST_AWK_COUNT=`echo '\n' | awk '{print $0}' | wc -c`
if [ "$TEST_AWK_COUNT" -eq 2 ]; then if [ "$TEST_AWK_COUNT" -eq 2 ]; then
@ -82,12 +83,27 @@ $FIREHOSE_TEST_HARNESS \
# Generate yaml from firehose output # Generate yaml from firehose output
. firehose-populated-commands.sh | awk ' . firehose-populated-commands.sh | awk '
BEGIN { BEGIN {
refseq_dir = "/humgen/gsa-hpprojects/GATK/data/Annotations/refseq/";
dbsnp_dir = "/humgen/gsa-hpprojects/GATK/data/";
dbsnps["Homo_sapiens_assembly18.fasta"] = dbsnp_dir "dbsnp_129_hg18.rod";
refseqs["Homo_sapiens_assembly18.fasta"] = refseq_dir "refGene-big-table-hg18.txt";
dbsnps["Homo_sapiens_assembly19.fasta"] = dbsnp_dir "dbsnp_132_b37.vcf";
refseqs["Homo_sapiens_assembly19.fasta"] = refseq_dir "refGene-big-table-hg19.txt";
printf "{" printf "{"
} }
{ {
if (NR == 1) { if (NR == 1) {
printf '"$PROJECT_YAML_TEMPLATE"' reference_part_count = split($1, reference_parts, "/")
printf "\n samples: [" reference_name = reference_parts[reference_part_count];
dbsnp = dbsnps[reference_name];
refseq = refseqs[reference_name];
printf '"$PROJECT_YAML_TEMPLATE"'
printf "\n samples: ["
} else { } else {
printf "," printf ","
} }