Merge pull request #317 from ncats/protein_disulf_hotfix

Protein disulf hotfix
ncats · Feb 1, 2024 · 6c4a1e0 · 6c4a1e0
2 parents 4fc5769 + 36b57fa
commit 6c4a1e0
Show file tree

Hide file tree

Showing 2 changed files with 37 additions and 0 deletions.
diff --git a/gsrs-module-substance-example/src/test/java/example/prot/ProtCalculationTest.java b/gsrs-module-substance-example/src/test/java/example/prot/ProtCalculationTest.java
@@ -646,6 +646,31 @@ public void proteinMolFormulaTest() {
                 "Formula must be the same when protein sequence includes lowercase codes");
     }
 
+    @Test
+    public void proteinMolFormulaDisulfidesTest() {
+        ProteinSubstance proteinSubstance = new ProteinSubstance();
+        Protein protein = new Protein();
+        Subunit subunit1= new  Subunit();
+        protein.subunits = new ArrayList<>();
+        protein.subunits.add(subunit1);
+        subunit1.sequence =
+                "CMMC";
+
+        DisulfideLink disulfideLink = new DisulfideLink();
+        List<Site> sitesDisulfide = new ArrayList<>();
+        sitesDisulfide.add(new Site(1, 1));
+        sitesDisulfide.add(new Site(1, 4));
+        disulfideLink.setSites(sitesDisulfide);
+        protein.setDisulfideLinks(Collections.singletonList(disulfideLink));
+        proteinSubstance.setProtein(protein);
+
+        Set<String> unknownResidues = new HashSet<>();
+
+        MolecularWeightAndFormulaContribution contribution=ProteinUtils.generateProteinWeightAndFormula(substanceRepository,
+                proteinSubstance, unknownResidues);
+        assertEquals("C16H28N4O5S4", contribution.getFormula());
+    }
+
     @Test
     public void aparaginePeptideMolFormulaTest() {
         ProteinSubstance proteinSubstance = new ProteinSubstance();

diff --git a/gsrs-module-substances-core/src/main/java/ix/ginas/utils/ProteinUtils.java b/gsrs-module-substances-core/src/main/java/ix/ginas/utils/ProteinUtils.java
@@ -324,6 +324,8 @@ public static MolecularWeightAndFormulaContribution generateProteinWeightAndForm
         log.trace(String.format("basic MW: %.2f; basic formula: %s; disulfideContribution: %.2f",
                 total, makeFormulaFromMap(formulaCounts), disulfideContribution));
         total -= disulfideContribution;
+        int numberOfHydrogensToRemove = getNumberOfHydrogensToRemoveForDisulfides(ps.protein);
+        log.trace("numberOfHydrogensToRemove: {}", numberOfHydrogensToRemove);
         Map<String, SingleThreadCounter> formulaMapWithContrib[] =new Map[1];
         formulaMapWithContrib[0]=formulaCounts;
         if (ps.hasModifications() && ps.modifications.structuralModifications.size() > 0) {
@@ -413,6 +415,10 @@ else if (contribution.getMw() > 0 || (contribution.getMwHigh() > 0 && contributi
             log.debug("no mods to consider");
         }
 
+        if( formulaMapWithContrib[0].containsKey("H")) {
+            log.trace("going to remove {} H atoms for disulfide bonds", numberOfHydrogensToRemove);
+            formulaMapWithContrib[0].get("H").decrement(numberOfHydrogensToRemove);
+        }
         log.trace(String.format("final total: %.2f; highTotal: %.2f; lowTotal: %.2f; highLimitTotal: %.2f; lowLimitTotal: %.2f", total,
                 highTotal, lowTotal, highLimitTotal, lowLimitTotal));
         result = new MolecularWeightAndFormulaContribution(total, ps.substanceClass.toString(), formulaMapWithContrib[0]);
@@ -738,6 +744,12 @@ public static double getDisulfideContribution(Protein protein) {
                 : 0.0d;
     }
 
+    public static int getNumberOfHydrogensToRemoveForDisulfides(Protein protein) {
+        return (protein != null && protein.getDisulfideLinks() != null)
+                ? 2 * protein.getDisulfideLinks().size()
+                : 0;
+    }
+
     public static Map<String,SingleThreadCounter> subtractFormulas(Map<String,SingleThreadCounter> minuend, Map<String,SingleThreadCounter> subtrahend){
         Map<String,SingleThreadCounter> difference = new HashMap<>();
         minuend.keySet().forEach(s->{