working better

alisman · Jan 2, 2025 · baf4516 · baf4516
1 parent 336515b
commit baf4516
Show file tree

Hide file tree

Showing 3 changed files with 55 additions and 1 deletion.
diff --git a/src/main/java/org/cbioportal/persistence/mybatisclickhouse/StudyViewMyBatisRepository.java b/src/main/java/org/cbioportal/persistence/mybatisclickhouse/StudyViewMyBatisRepository.java
@@ -384,6 +384,8 @@ public HashMap<String, AlterationCountByGene> getAlterationEnrichmentCounts(List
                 n.getValue().setNumberOfProfiledCases(
                     geneCount.get(n.getKey()).getNumberOfProfiledCases()
                 );
+            } else {
+                n.getValue().setNumberOfProfiledCases(0);
             }
         });
 

diff --git a/src/main/java/org/cbioportal/service/util/AlterationEnrichmentUtil.java b/src/main/java/org/cbioportal/service/util/AlterationEnrichmentUtil.java
@@ -41,7 +41,35 @@ public class AlterationEnrichmentUtil<T extends AlterationCountBase> {
     private GenePanelService genePanelService;
     @Autowired
     private MolecularProfileService molecularProfileService;
-
+
+    public double calculatePValue(List<CountSummary> counts){
+        double pValue;
+        if (counts.size() == 2) {
+            int alteredInNoneCount = counts.get(1).getProfiledCount() - counts.get(1).getAlteredCount();
+            int alteredOnlyInQueryGenesCount = counts.get(0).getProfiledCount()
+                - counts.get(0).getAlteredCount();
+
+            pValue = fisherExactTestCalculator.getTwoTailedPValue(alteredInNoneCount,
+                counts.get(1).getAlteredCount(), alteredOnlyInQueryGenesCount,
+                counts.get(0).getAlteredCount());
+        } else {
+
+            long[][] array = counts.stream().map(count -> {
+                return new long[]{count.getAlteredCount(),
+                    count.getProfiledCount() - count.getAlteredCount()};
+            }).toArray(long[][]::new);
+
+            ChiSquareTest chiSquareTest = new ChiSquareTest();
+            pValue = chiSquareTest.chiSquareTest(array);
+
+            // set p-value to 1 when the cases in all groups are altered
+            if (Double.isNaN(pValue)) {
+                pValue = 1;
+            }
+        }
+        return pValue;
+    }
+
     public List<AlterationEnrichment> createAlterationEnrichments(Map<String, Pair<List<AlterationCountByGene>, Long>> mutationCountsbyGroup) {
 
         Map<String, Map<Integer, AlterationCountByGene>> mutationCountsbyEntrezGeneIdAndGroup = mutationCountsbyGroup

diff --git a/src/main/java/org/cbioportal/web/columnar/EnrichmentsColumnStoreController.java b/src/main/java/org/cbioportal/web/columnar/EnrichmentsColumnStoreController.java
@@ -9,6 +9,7 @@
 import io.swagger.v3.oas.annotations.media.Schema;
 import io.swagger.v3.oas.annotations.responses.ApiResponse;
 import jakarta.validation.Valid;
+import org.apache.commons.math3.stat.inference.ChiSquareTest;
 import org.cbioportal.model.AlterationCountByGene;
 import org.cbioportal.model.AlterationEnrichment;
 import org.cbioportal.model.AlterationFilter;
@@ -40,6 +41,7 @@
 import org.cbioportal.service.ViolinPlotService;
 import org.cbioportal.service.exception.MolecularProfileNotFoundException;
 import org.cbioportal.service.exception.StudyNotFoundException;
+import org.cbioportal.service.util.AlterationEnrichmentUtil;
 import org.cbioportal.service.util.CustomDataSession;
 import org.cbioportal.utils.config.annotation.ConditionalOnProperty;
 import org.cbioportal.web.columnar.util.CustomDataFilterUtil;
@@ -151,6 +153,8 @@ public ResponseEntity<List<AlterationEnrichment>> fetchAlterationEnrichmentsNew(
         @Parameter(required = true, description = "List of groups containing sample identifiers and list of Alteration Types")
         @Valid @RequestBody(required = false) MolecularProfileCasesGroupAndAlterationTypeFilter groupsAndAlterationTypes) throws MolecularProfileNotFoundException {
 
+
+        var util = new AlterationEnrichmentUtil();
 
         // this needs to be the response type
         //List<AlterationEnrichment>
@@ -206,9 +210,29 @@ public ResponseEntity<List<AlterationEnrichment>> fetchAlterationEnrichmentsNew(
              }).collect(Collectors.toList());
              enrichment.setCounts(counts);
 
+             var filteredCounts = counts.stream().filter(groupCasesCount -> groupCasesCount.getProfiledCount() > 0)
+                 .collect(Collectors.toList());
+
+             // groups where number of altered cases is greater than profiled cases.
+             // This is a temporary fix for https://github.com/cBioPortal/cbioportal/issues/7274
+             // and https://github.com/cBioPortal/cbioportal/issues/7418
+             long invalidDataGroups = filteredCounts
+                 .stream()
+                 .filter(groupCasesCount -> groupCasesCount.getAlteredCount() > groupCasesCount.getProfiledCount())
+                 .count();
+
+             // calculate p-value only if more than one group have profile cases count
+             // greater than 0
+             if (filteredCounts.size() > 1 && invalidDataGroups == 0) {
+                 enrichment.setpValue(BigDecimal.valueOf(util.calculatePValue(counts)));
+             }
+
              return enrichment;
 
          }).collect(Collectors.toList());
+
+
+
 
 
 //        List<AlterationEnrichment> alterationEnrichments = alterationEnrichmentService.getAlterationEnrichments(