Skip to content

Commit

Permalink
changed IS_GERMLINE to SV_STATUS
Browse files Browse the repository at this point in the history
  • Loading branch information
@mandawilson
mandawilson committed Feb 9, 2022
1 parent e55e602 commit 4fe21b3
Show file tree
Hide file tree
Showing 18 changed files with 143 additions and 140 deletions.
11 changes: 5 additions & 6 deletions core/src/main/java/org/mskcc/cbio/portal/dao/DaoStructuralVariant.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -81,7 +81,7 @@ public static void addStructuralVariantToBulkLoader(StructuralVariant structural
"LENGTH",
"COMMENTS",
"EXTERNAL_ANNOTATION",
"IS_GERMLINE",
"SV_STATUS",
};
bl.setFieldNames(fieldNames);

Expand All @@ -90,13 +90,13 @@ public static void addStructuralVariantToBulkLoader(StructuralVariant structural
Long.toString(structuralVariant.getInternalId()),
Integer.toString(structuralVariant.getGeneticProfileId()),
Integer.toString(structuralVariant.getSampleIdInternal()),
Long.toString(structuralVariant.getSite1EntrezGeneId()),
structuralVariant.getSite1EntrezGeneId() == null ? null : Long.toString(structuralVariant.getSite1EntrezGeneId()),
structuralVariant.getSite1EnsemblTranscriptId(),
Integer.toString(structuralVariant.getSite1Exon()),
structuralVariant.getSite1Chromosome(),
Integer.toString(structuralVariant.getSite1Position()),
structuralVariant.getSite1Description(),
Long.toString(structuralVariant.getSite2EntrezGeneId()),
structuralVariant.getSite2EntrezGeneId() == null ? null : Long.toString(structuralVariant.getSite2EntrezGeneId()),
structuralVariant.getSite2EnsemblTranscriptId(),
Integer.toString(structuralVariant.getSite2Exon()),
structuralVariant.getSite2Chromosome(),
Expand All @@ -123,8 +123,7 @@ public static void addStructuralVariantToBulkLoader(StructuralVariant structural
Integer.toString(structuralVariant.getLength()),
structuralVariant.getComments(),
structuralVariant.getExternalAnnotation(),
Integer.toString(structuralVariant.getIsGermline()?1:0));
//TODO: fix this ... the value must be converted to an integer
structuralVariant.getSvStatus());

if ((structuralVariant.getDriverFilter() != null
&& !structuralVariant.getDriverFilter().isEmpty()
Expand Down Expand Up @@ -240,7 +239,7 @@ private static StructuralVariant extractStructuralVariant(ResultSet rs) throws S
structuralVariant.setDriverFilterAnn(rs.getString("DRIVER_FILTER_ANNOTATION"));
structuralVariant.setDriverTiersFilter(rs.getString("DRIVER_TIERS_FILTER"));
structuralVariant.setDriverTiersFilterAnn(rs.getString("DRIVER_TIERS_FILTER_ANNOTATION"));
structuralVariant.setIsGermline(rs.getBoolean("IS_GERMLINE"));
structuralVariant.setSvStatus(rs.getString("SV_STATUS"));
return structuralVariant;
}
}
Expand Down
10 changes: 5 additions & 5 deletions core/src/main/java/org/mskcc/cbio/portal/model/StructuralVariant.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -69,7 +69,7 @@ public class StructuralVariant {
private String driverFilterAnn;
private String driverTiersFilter;
private String driverTiersFilterAnn;
private Boolean isGermline;
private String svStatus;

public long getInternalId() {
return internalId;
Expand Down Expand Up @@ -335,10 +335,10 @@ public String getDriverTiersFilterAnn() {
public void setDriverTiersFilterAnn(String driverTiersFilterAnn) {
this.driverTiersFilterAnn = driverTiersFilterAnn;
}
public Boolean getIsGermline() {
return isGermline;
public String getSvStatus() {
return svStatus;
}
public void setIsGermline(Boolean isGermline) {
this.isGermline = isGermline;
public void setSvStatus(String svStatus) {
this.svStatus = svStatus;
}
}
23 changes: 12 additions & 11 deletions core/src/main/java/org/mskcc/cbio/portal/scripts/ImportStructuralVariantData.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -111,22 +111,23 @@ public void importData() throws IOException, DaoException {
CanonicalGene site2CanonicalGene = setCanonicalGene(site2EntrezGeneId, site2HugoSymbol, daoGene);

// If neither of the genes is recognized, skip the line
if(site1CanonicalGene == null) {
ProgressMonitor.logWarning("Gene not found: " + site1HugoSymbol + " ["
+ site1EntrezGeneId + "]. Ignoring it "
+ "and all fusion data associated with it!");
} else if (site2CanonicalGene == null) {
ProgressMonitor.logWarning("Gene not found: " + site2HugoSymbol + " ["
if(site1CanonicalGene == null && site2CanonicalGene == null) {
ProgressMonitor.logWarning("Could not find gene 1: " + site1HugoSymbol + " [" + site1EntrezGeneId
+ "] or gene 2: " + site2HugoSymbol + " ["
+ site2EntrezGeneId + "]. Ignoring it "
+ "and all fusion data associated with it!");
// If both genes are recognized, continue
+ "and all SV data associated with it!");
// If at least one gene is recognized, continue
} else {
// Save the Entrez Gene Id if it was not saved before
if (site1EntrezGeneId == TabDelimitedFileUtil.NA_LONG) {
if (site1EntrezGeneId == TabDelimitedFileUtil.NA_LONG && site1CanonicalGene != null) {
structuralVariant.setSite1EntrezGeneId(site1CanonicalGene.getEntrezGeneId());
} else if (site1EntrezGeneId == TabDelimitedFileUtil.NA_LONG) {
structuralVariant.setSite1EntrezGeneId(null); // we want this to be null in the database, not NA_LONG
}
if (site2EntrezGeneId == TabDelimitedFileUtil.NA_LONG) {
if (site2EntrezGeneId == TabDelimitedFileUtil.NA_LONG && site2CanonicalGene != null) {
structuralVariant.setSite2EntrezGeneId(site2CanonicalGene.getEntrezGeneId());
} else if (site2EntrezGeneId == TabDelimitedFileUtil.NA_LONG) {
structuralVariant.setSite2EntrezGeneId(null); // we want this to be null in the database, not NA_LONG
}
// Add structural variant
DaoStructuralVariant.addStructuralVariantToBulkLoader(structuralVariant);
Expand Down Expand Up @@ -160,7 +161,7 @@ private CanonicalGene setCanonicalGene(long siteEntrezGeneId, String siteHugoSym
}

// If no gene can be found based on Entrez Gene ID, try Symbol.
if (siteCanonicalGene == null) {
if (siteCanonicalGene == null && !TabDelimitedFileUtil.NA_STRING.equals(siteHugoSymbol)) {
siteCanonicalGene = daoGene.getNonAmbiguousGene(siteHugoSymbol, true);
}

Expand Down
14 changes: 8 additions & 6 deletions core/src/main/java/org/mskcc/cbio/portal/util/StructuralVariantUtil.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -83,8 +83,8 @@ public class StructuralVariantUtil {
public static final String DRIVER_FILTER_ANNOTATION = "cbp_driver_annotation";
public static final String DRIVER_TIERS_FILTER = "cbp_driver_tiers";
public static final String DRIVER_TIERS_FILTER_ANNOTATION = "cbp_driver_tiers_annotation";
public static final String IS_GERMLINE = "is_germline";

public static final String SV_STATUS = "Sv_Status";
public StructuralVariantUtil(){}

public StructuralVariantUtil(String line) {
Expand Down Expand Up @@ -140,7 +140,10 @@ public StructuralVariant parseStructuralVariantRecord(String[] parts) {
structuralVariant.setDriverFilterAnn(TabDelimitedFileUtil.getPartString(getColumnIndex(StructuralVariantUtil.DRIVER_FILTER_ANNOTATION), parts));
structuralVariant.setDriverTiersFilter(TabDelimitedFileUtil.getPartString(getColumnIndex(StructuralVariantUtil.DRIVER_TIERS_FILTER), parts));
structuralVariant.setDriverTiersFilterAnn(TabDelimitedFileUtil.getPartString(getColumnIndex(StructuralVariantUtil.DRIVER_TIERS_FILTER_ANNOTATION), parts));
structuralVariant.setIsGermline(TabDelimitedFileUtil.getPartInt(getColumnIndex(StructuralVariantUtil.IS_GERMLINE), parts) != 0);
structuralVariant.setSvStatus(TabDelimitedFileUtil.getPartString(getColumnIndex(StructuralVariantUtil.SV_STATUS), parts));
if (TabDelimitedFileUtil.NA_STRING.equals(structuralVariant.getSvStatus())) {
structuralVariant.setSvStatus(null); // we want to use the database default
}
return structuralVariant;
}

Expand All @@ -162,7 +165,7 @@ public int getColumnIndex(String colName) {
* If a structural variant record has a mix of defined and missing values for Site 1 or Site 2
* Ensembl transcript IDs and/or exons then the structural variant record will not be imported.
*
* Example (assuming that site 1 and site 2 hugo symbol and/or entrez id are present):
* Example (assuming that site 1 or site 2 hugo symbol and/or entrez id are present):
*
* Valid Record:
* Site 1 Transcript: EST0000024958
Expand Down Expand Up @@ -195,8 +198,7 @@ public Boolean hasRequiredStructuralVariantFields(StructuralVariant record) {
record.getSite1Exon() != -1 &&
record.getSite2Exon() != -1);
return ( hasNoEnsemblExonValues || hasAllEnsemblExonValues ) &&
(record.getSite1EntrezGeneId() != Long.MIN_VALUE || !record.getSite1HugoSymbol().equalsIgnoreCase(TabDelimitedFileUtil.NA_STRING)) &&
(record.getSite2EntrezGeneId() != Long.MIN_VALUE || !record.getSite2HugoSymbol().equalsIgnoreCase(TabDelimitedFileUtil.NA_STRING));
(record.getSite1EntrezGeneId() != Long.MIN_VALUE || !record.getSite1HugoSymbol().equalsIgnoreCase(TabDelimitedFileUtil.NA_STRING) || record.getSite2EntrezGeneId() != Long.MIN_VALUE || !record.getSite2HugoSymbol().equalsIgnoreCase(TabDelimitedFileUtil.NA_STRING));
}

}
6 changes: 3 additions & 3 deletions core/src/test/resources/data_structural_variants.txt
View file
Original file line number Diff line number Diff line change
@@ -1,3 +1,3 @@
Sample_ID Site1_Entrez_Gene_Id Site1_Hugo_Symbol Site1_Ensembl_Transcript_Id Site1_Exon Site1_Chromosome Site1_Position Site1_Description Site2_Entrez_Gene_Id Site2_Hugo_Symbol Site2_Ensembl_Transcript_Id Site2_Exon Site2_Chromosome Site2_Position Site2_Description Site2_Effect_On_Frame NCBI_Build DNA_Support RNA_Support Normal_Read_Count Tumor_Read_Count Normal_Variant_Count Tumor_Variant_Count Normal_Paired_End_Read_Count Tumor_Paired_End_Read_Count Normal_Split_Read_Count Tumor_Split_Read_Count Annotation Breakpoint_Type Center Connection_Type Event_Info Class Length Comments External_Annotation cbp_driver cbp_driver_annotation cbp_driver_tiers cbp_driver_tiers_annotation is_germline
TCGA-A1-A0SB-01 NA KIAA1549 ENST00000242365 15 7 138536968 KIAA1549-BRAF.K16B10.COSF509_1 NA BRAF ENST00000288602 10 7 140482957 KIAA1549-BRAF.K16B10.COSF509_2 NA GRCh37 no yes NA 1000 NA 900 NA NA NA NA KIAA1549-BRAF.K16B10.COSF509 NA NA NA Fusion NA NA Gain-of-Function COSMIC:COSF509 NA NA NA NA 0
TCGA-A1-A0SD-01 NA EML4 ENST00000318522 13 2 42522656 EML4-ALK.E13A20.AB462411_1 NA ALK ENST00000389048 20 2 29446335 EML4-ALK.E13A20.AB462411_2 NA GRCh37 no yes NA 1006 NA 300 NA NA NA NA EML4-ALK.E13A20 NA NA NA Fusion NA NA Gain-of-Function GENBANK:AB462411 NA NA NA NA 0
Sample_ID Site1_Entrez_Gene_Id Site1_Hugo_Symbol Site1_Ensembl_Transcript_Id Site1_Exon Site1_Chromosome Site1_Position Site1_Description Site2_Entrez_Gene_Id Site2_Hugo_Symbol Site2_Ensembl_Transcript_Id Site2_Exon Site2_Chromosome Site2_Position Site2_Description Site2_Effect_On_Frame NCBI_Build DNA_Support RNA_Support Normal_Read_Count Tumor_Read_Count Normal_Variant_Count Tumor_Variant_Count Normal_Paired_End_Read_Count Tumor_Paired_End_Read_Count Normal_Split_Read_Count Tumor_Split_Read_Count Annotation Breakpoint_Type Center Connection_Type Event_Info Class Length Comments External_Annotation cbp_driver cbp_driver_annotation cbp_driver_tiers cbp_driver_tiers_annotation Sv_Status
TCGA-A1-A0SB-01 NA KIAA1549 ENST00000242365 15 7 138536968 KIAA1549-BRAF.K16B10.COSF509_1 NA BRAF ENST00000288602 10 7 140482957 KIAA1549-BRAF.K16B10.COSF509_2 NA GRCh37 no yes NA 1000 NA 900 NA NA NA NA KIAA1549-BRAF.K16B10.COSF509 NA NA NA Fusion NA NA Gain-of-Function COSMIC:COSF509 NA NA NA NA SOMATIC
TCGA-A1-A0SD-01 NA EML4 ENST00000318522 13 2 42522656 EML4-ALK.E13A20.AB462411_1 NA ALK ENST00000389048 20 2 29446335 EML4-ALK.E13A20.AB462411_2 NA GRCh37 no yes NA 1006 NA 300 NA NA NA NA EML4-ALK.E13A20 NA NA NA Fusion NA NA Gain-of-Function GENBANK:AB462411 NA NA NA NA SOMATIC
16 changes: 8 additions & 8 deletions core/src/test/scripts/test_data/data_structural_variants_exon_not_in_transcript.txt
View file
Original file line number Diff line number Diff line change
@@ -1,8 +1,8 @@
Sample_ID Site1_Entrez_Gene_Id Site1_Hugo_Symbol Site1_Ensembl_Transcript_Id Site1_Exon Site1_Chromosome Site1_Position Site1_Description Site2_Entrez_Gene_Id Site2_Hugo_Symbol Site2_Ensembl_Transcript_Id Site2_Exon Site2_Chromosome Site2_Position Site2_Description Site2_Effect_On_Frame NCBI_Build DNA_Support RNA_Support Normal_Read_Count Tumor_Read_Count Normal_Variant_Count Tumor_Variant_Count Normal_Paired_End_Read_Count Tumor_Paired_End_Read_Count Normal_Split_Read_Count Tumor_Split_Read_Count Annotation Breakpoint_Type Center Connection_Type Event_Info Class Length Comments External_Annotation cbp_driver cbp_driver_annotation cbp_driver_tiers cbp_driver_tiers_annotation is_germline
TCGA-A2-A04P-01 NA KIAA1549 ENST00000242365 1500 7 138536968 KIAA1549-BRAF.K16B10.COSF509_1 NA BRAF ENST00000288602 10 7 140482957 KIAA1549-BRAF.K16B10.COSF509_2 NA GRCh37 no yes NA 1000 NA 900 NA NA NA NA KIAA1549-BRAF.K16B10.COSF509 NA NA NA Fusion NA NA Gain-of-Function COSMIC:COSF509 NA NA NA NA 0
TCGA-A2-A04P-01 NA NCOA4 ENST00000344348 7 10 51582939 NCOA4-RET.N7R12_1 NA RET ENST00000340058 12 10 43612031 NCOA4-RET.N7R12_2 NA GRCh37 no yes NA 1001 NA 800 NA NA NA NA NCOA4-RET.N7R1 NA NA NA Fusion NA NA Gain-of-Function NA NA NA NA NA 0
TCGA-A2-A04P-01 NA EML4 ENST00000318522 6 2 42492091 EML4-ALK.E6bA20.AB374362_1 NA ALK ENST00000389048 2000 2 29446394 EML4-ALK.E6bA20.AB374362_2 NA GRCh37 no yes NA 1002 NA 700 NA NA NA NA EML4-ALK.E6bA20.AB374362 NA NA NA Fusion NA NA Gain-of-Function GENBANK:AB374362 NA NA NA NA 0
TCGA-A2-A04P-01 NA TMPRSS2 ENST00000332149 1 21 42880007 TMPRSS2-ERG.T1E2.COSF23.1_1 NA ERG ENST00000442448 2 21 39956869 TMPRSS2-ERG.T1E2.COSF23.1_2 NA GRCh37 no yes NA 1003 NA 600 NA NA NA NA TMPRSS2-ERG.T1E2.COSF23.1 NA NA NA Fusion NA NA Gain-of-Function COSMIC:COSF23 NA NA NA NA 0
TCGA-A2-A04P-01 NA EGFR ENST00000275493 1 7 55087058 EGFR-EGFR.E1E8.DelPositive.1_1 NA EGFR ENST00000275493 8 7 55223522 EGFR-EGFR.E1E8.DelPositive.1_2 NA GRCh37 no yes NA 1004 NA 500 NA NA NA NA EGFR-EGFR.E1E8.DelPositive NA NA NA Fusion NA NA NA NA NA NA NA NA 0
TCGA-A2-A04P-01 NA ALK ENST00000389048 11 2 29497964 ALK-PTPN3.A11P3_1 NA PTPN3 ENST00000374541 3 9 112219679 ALK-PTPN3.A11P3_2 NA GRCh37 no yes NA 1005 NA 400 NA NA NA NA ALK-PTPN3.A11P3 NA NA NA Fusion NA NA NA NA NA NA NA NA 0
TCGA-A1-A0SB-01 NA EML4 ENST00000318522 13 2 42522656 EML4-ALK.E13A20.AB462411_1 NA ALK ENST00000389048 20 2 29446335 EML4-ALK.E13A20.AB462411_2 NA GRCh37 no yes NA 1006 NA 300 NA NA NA NA EML4-ALK.E13A20 NA NA NA Fusion NA NA Gain-of-Function GENBANK:AB462411 NA NA NA NA 0
Sample_ID Site1_Entrez_Gene_Id Site1_Hugo_Symbol Site1_Ensembl_Transcript_Id Site1_Exon Site1_Chromosome Site1_Position Site1_Description Site2_Entrez_Gene_Id Site2_Hugo_Symbol Site2_Ensembl_Transcript_Id Site2_Exon Site2_Chromosome Site2_Position Site2_Description Site2_Effect_On_Frame NCBI_Build DNA_Support RNA_Support Normal_Read_Count Tumor_Read_Count Normal_Variant_Count Tumor_Variant_Count Normal_Paired_End_Read_Count Tumor_Paired_End_Read_Count Normal_Split_Read_Count Tumor_Split_Read_Count Annotation Breakpoint_Type Center Connection_Type Event_Info Class Length Comments External_Annotation cbp_driver cbp_driver_annotation cbp_driver_tiers cbp_driver_tiers_annotation Sv_Status
TCGA-A2-A04P-01 NA KIAA1549 ENST00000242365 1500 7 138536968 KIAA1549-BRAF.K16B10.COSF509_1 NA BRAF ENST00000288602 10 7 140482957 KIAA1549-BRAF.K16B10.COSF509_2 NA GRCh37 no yes NA 1000 NA 900 NA NA NA NA KIAA1549-BRAF.K16B10.COSF509 NA NA NA Fusion NA NA Gain-of-Function COSMIC:COSF509 NA NA NA NA SOMATIC
TCGA-A2-A04P-01 NA NCOA4 ENST00000344348 7 10 51582939 NCOA4-RET.N7R12_1 NA RET ENST00000340058 12 10 43612031 NCOA4-RET.N7R12_2 NA GRCh37 no yes NA 1001 NA 800 NA NA NA NA NCOA4-RET.N7R1 NA NA NA Fusion NA NA Gain-of-Function NA NA NA NA NA SOMATIC
TCGA-A2-A04P-01 NA EML4 ENST00000318522 6 2 42492091 EML4-ALK.E6bA20.AB374362_1 NA ALK ENST00000389048 2000 2 29446394 EML4-ALK.E6bA20.AB374362_2 NA GRCh37 no yes NA 1002 NA 700 NA NA NA NA EML4-ALK.E6bA20.AB374362 NA NA NA Fusion NA NA Gain-of-Function GENBANK:AB374362 NA NA NA NA SOMATIC
TCGA-A2-A04P-01 NA TMPRSS2 ENST00000332149 1 21 42880007 TMPRSS2-ERG.T1E2.COSF23.1_1 NA ERG ENST00000442448 2 21 39956869 TMPRSS2-ERG.T1E2.COSF23.1_2 NA GRCh37 no yes NA 1003 NA 600 NA NA NA NA TMPRSS2-ERG.T1E2.COSF23.1 NA NA NA Fusion NA NA Gain-of-Function COSMIC:COSF23 NA NA NA NA SOMATIC
TCGA-A2-A04P-01 NA EGFR ENST00000275493 1 7 55087058 EGFR-EGFR.E1E8.DelPositive.1_1 NA EGFR ENST00000275493 8 7 55223522 EGFR-EGFR.E1E8.DelPositive.1_2 NA GRCh37 no yes NA 1004 NA 500 NA NA NA NA EGFR-EGFR.E1E8.DelPositive NA NA NA Fusion NA NA NA NA NA NA NA NA SOMATIC
TCGA-A2-A04P-01 NA ALK ENST00000389048 11 2 29497964 ALK-PTPN3.A11P3_1 NA PTPN3 ENST00000374541 3 9 112219679 ALK-PTPN3.A11P3_2 NA GRCh37 no yes NA 1005 NA 400 NA NA NA NA ALK-PTPN3.A11P3 NA NA NA Fusion NA NA NA NA NA NA NA NA SOMATIC
TCGA-A1-A0SB-01 NA EML4 ENST00000318522 13 2 42522656 EML4-ALK.E13A20.AB462411_1 NA ALK ENST00000389048 20 2 29446335 EML4-ALK.E13A20.AB462411_2 NA GRCh37 no yes NA 1006 NA 300 NA NA NA NA EML4-ALK.E13A20 NA NA NA Fusion NA NA Gain-of-Function GENBANK:AB462411 NA NA NA NA SOMATIC
Loading

0 comments on commit 4fe21b3

Please sign in to comment.