Update gene constraints to use gnomad v4.0 data

Correct spelling of GeneConstraints.geneContraint() to geneConstraint()

exomiser-core/src/main/java/org/monarchinitiative/exomiser/core/analysis/util/GeneConstraint.java

@@ -41,47 +41,7 @@
 /**
  * @since 13.1.0
  */
-public class GeneConstraint {
-
-    private final String geneSymbol;
-    private final String transcriptId;
-    private final double pLI;
-    private final double loeuf;
-    private final double loeufLower;
-    private final double loeufUpper;
-
-    public GeneConstraint(String geneSymbol, String transcriptId, double pLI, double loeuf, double loeufLower, double loeufUpper) {
-        this.geneSymbol = geneSymbol;
-        this.transcriptId = transcriptId;
-        this.pLI = pLI;
-        this.loeuf = loeuf;
-        this.loeufLower = loeufLower;
-        this.loeufUpper = loeufUpper;
-    }
-
-    public String geneSymbol() {
-        return geneSymbol;
-    }
-
-    public String transcriptId() {
-        return transcriptId;
-    }
-
-    public double pLI() {
-        return pLI;
-    }
-
-    public double loeuf() {
-        return loeuf;
-    }
-
-    public double loeufLower() {
-        return loeufLower;
-    }
-
-    public double loeufUpper() {
-        return loeufUpper;
-    }
+public record GeneConstraint(String geneSymbol, String transcriptId, double pLI, double loeuf, double loeufLower, double loeufUpper, double missenseZ, double synonymousZ) {
 
     /**
      * For the interpretation of Mendelian diseases cases, we suggest using the upper bound of the oe CI < 0.35 as a threshold if needed.
@@ -92,31 +52,7 @@ public boolean isLossOfFunctionIntolerant() {
         //gnomAD suggest using a loeufUpper < 0.35;
         // However varsome suggests a relaxed metric has better recall due to lower FP.
         // https://varsome.com/about/resources/acmg-implementation/#pvs1
-        return loeuf < 0.7635;
+        return loeuf < 0.7555;
     }
 
-    @Override
-    public boolean equals(Object o) {
-        if (this == o) return true;
-        if (o == null || getClass() != o.getClass()) return false;
-        GeneConstraint that = (GeneConstraint) o;
-        return Double.compare(that.pLI, pLI) == 0 && Double.compare(that.loeuf, loeuf) == 0 && Double.compare(that.loeufLower, loeufLower) == 0 && Double.compare(that.loeufUpper, loeufUpper) == 0 && geneSymbol.equals(that.geneSymbol) && transcriptId.equals(that.transcriptId);
-    }
-
-    @Override
-    public int hashCode() {
-        return Objects.hash(geneSymbol, transcriptId, pLI, loeuf, loeufLower, loeufUpper);
-    }
-
-    @Override
-    public String toString() {
-        return "GeneContraint{" +
-                "geneSymbol='" + geneSymbol + '\'' +
-                ", transcriptId='" + transcriptId + '\'' +
-                ", pLI=" + pLI +
-                ", loeuf=" + loeuf +
-                ", loeufLower=" + loeufLower +
-                ", loeufUpper=" + loeufUpper +
-                '}';
-    }
 }

exomiser-core/src/main/java/org/monarchinitiative/exomiser/core/analysis/util/GeneConstraints.java

@@ -25,24 +25,25 @@
 import java.io.IOException;
 import java.io.InputStream;
 import java.io.InputStreamReader;
-import java.util.Collections;
-import java.util.LinkedHashMap;
-import java.util.Map;
-import java.util.Optional;
+import java.util.*;
 
 /**
  * @since 13.1.0
  */
 public class GeneConstraints {
 
-    private static final Map<String, GeneConstraint> geneConstraints = GnomadGeneConstraintParser.readGeneConstraints("gnomad.v2.1.1.gene-constraints.tsv");
+    private static final Map<String, GeneConstraint> GENE_CONSTRAINTS = GnomadGeneConstraintParser.readGeneConstraints("gnomad.v4.0.gene-constraints.tsv");
 
     private GeneConstraints() {
     }
 
     @Nullable
-    public static GeneConstraint geneContraint(String geneSymbol) {
-        return geneConstraints.get(geneSymbol);
+    public static GeneConstraint geneConstraint(String geneSymbol) {
+        return GENE_CONSTRAINTS.get(geneSymbol);
+    }
+
+    public static Collection<GeneConstraint> geneConstraints() {
+        return GENE_CONSTRAINTS.values();
     }
 
     private static class GnomadGeneConstraintParser {
@@ -53,9 +54,9 @@ private static Map<String, GeneConstraint> readGeneConstraints(String constraint
             try (BufferedReader bufferedReader = new BufferedReader(new InputStreamReader(getResourceAsStream(constraintsFile)))) {
                 for (String line; (line = bufferedReader.readLine()) != null; ) {
                     if (!line.startsWith("#")) {
-                        GeneConstraint geneContraint = parseGeneConstraint(line);
-                        if (!Double.isNaN(geneContraint.loeufUpper())) {
-                            geneConstraints.put(geneContraint.geneSymbol(), geneContraint);
+                        GeneConstraint geneConstraint = parseGeneConstraint(line);
+                        if (geneConstraint != null) {
+                            geneConstraints.put(geneConstraint.geneSymbol(), geneConstraint);
                         }
                     }
                 }
@@ -66,14 +67,21 @@ private static Map<String, GeneConstraint> readGeneConstraints(String constraint
         }
 
         private static GeneConstraint parseGeneConstraint(String line) {
+            // gene	transcript	mane_select	lof.oe	lof.pLI	lof.oe_ci.lower	lof.oe_ci.upper	mis.z_score	syn.z_score
             String[] fields = line.split("\t");
             String geneSymbol = fields[0];
             String transcriptId = fields[1];
-            double pLI = parseDouble(fields[2]);
+            boolean isManeSelect = Boolean.parseBoolean(fields[2]);
+            if (geneSymbol.equals("NA") || !isManeSelect) {
+                return null;
+            }
+            double pLI = parseDouble(fields[4]);
             double loeuf = parseDouble(fields[3]);
-            double loeufLower = parseDouble(fields[4]);
-            double loeufUpper = parseDouble(fields[5]);
-            return new GeneConstraint(geneSymbol, transcriptId, pLI, loeuf, loeufLower, loeufUpper);
+            double loeufLower = parseDouble(fields[5]);
+            double loeufUpper = parseDouble(fields[6]);
+            double missenseZ = parseDouble(fields[7]);
+            double synonymousZ = parseDouble(fields[8]);
+            return new GeneConstraint(geneSymbol, transcriptId, pLI, loeuf, loeufLower, loeufUpper, missenseZ, synonymousZ);
         }
 
         private static double parseDouble(String field) {

exomiser-core/src/main/java/org/monarchinitiative/exomiser/core/analysis/util/acmg/Acmg2015EvidenceAssigner.java

@@ -210,7 +210,7 @@ private void assignPVS1(AcmgEvidence.Builder acmgEvidenceBuilder, VariantEvaluat
         // •  Use caution with splice variants that are predicted to lead to exon skipping but leave the remainder of the protein intact
         // •  Use caution in the presence of multiple transcripts
 
-        GeneConstraint geneContraint = GeneConstraints.geneContraint(variantEvaluation.getGeneSymbol());
+        GeneConstraint geneContraint = GeneConstraints.geneConstraint(variantEvaluation.getGeneSymbol());
         // Should this be using the hasCompatibleDiseaseMatches variable?
         boolean inGeneWithKnownDiseaseAssociations = !knownDiseases.isEmpty();
         if (inGeneWithKnownDiseaseAssociations && isLossOfFunctionEffect(variantEvaluation.getVariantEffect())

exomiser-core/src/main/java/org/monarchinitiative/exomiser/core/writers/TsvVariantAllMoiResultsWriter.java

@@ -165,7 +165,7 @@ private List<Object> buildVariantRecord(int rank, VariantEvaluation ve, GeneScor
         fields.add(clinVarData.getVariationId());
         fields.add(clinVarData.getPrimaryInterpretation());
         fields.add(clinVarData.starRating());
-        GeneConstraint geneConstraint = GeneConstraints.geneContraint(geneIdentifier.getGeneSymbol());
+        GeneConstraint geneConstraint = GeneConstraints.geneConstraint(geneIdentifier.getGeneSymbol());
         fields.add(geneConstraint == null ? "" : geneConstraint.loeuf());
         fields.add(geneConstraint == null ? "" : geneConstraint.loeufLower());
         fields.add(geneConstraint == null ? "" : geneConstraint.loeufUpper());

exomiser-core/src/main/java/org/monarchinitiative/exomiser/core/writers/TsvVariantResultsWriter.java

@@ -167,7 +167,7 @@ private List<Object> buildVariantRecord(int rank, VariantEvaluation ve, GeneScor
         fields.add(clinVarData.getVariationId());
         fields.add(clinVarData.getPrimaryInterpretation());
         fields.add(clinVarData.starRating());
-        GeneConstraint geneConstraint = GeneConstraints.geneContraint(geneIdentifier.getGeneSymbol());
+        GeneConstraint geneConstraint = GeneConstraints.geneConstraint(geneIdentifier.getGeneSymbol());
         fields.add(geneConstraint == null ? "" : geneConstraint.loeuf());
         fields.add(geneConstraint == null ? "" : geneConstraint.loeufLower());
         fields.add(geneConstraint == null ? "" : geneConstraint.loeufUpper());