Add configurable ProForma residue validation

.github/workflows/dotnet.yml

@@ -17,7 +17,7 @@ jobs:
     - name: Setup .NET
       uses: actions/setup-dotnet@v4
       with:
-        dotnet-version: 6.x
+        dotnet-version: 8.x
     - name: Restore dependencies
       run: dotnet restore
     - name: Build

.github/workflows/release.yml

@@ -14,7 +14,7 @@ jobs:
       - name: Setup .NET
         uses: actions/setup-dotnet@v3
         with:
-          dotnet-version: 6.x
+          dotnet-version: 8.x
 
       - name: Verify commit exists in origin/master
         run: |

src/TopDownProteomics.Benchmarks/TopDownProteomics.Benchmarks.csproj

@@ -3,7 +3,7 @@
     <OutputType>Exe</OutputType>
     <RootNamespace>TopDownProteomics.Benchmarks</RootNamespace>
     <AssemblyName>TopDownProteomics.Benchmarks</AssemblyName>
-    <TargetFramework>net6.0</TargetFramework>
+    <TargetFramework>net8.0</TargetFramework>
     <Authors>Consortium for Top-Down Proteomics</Authors>
   </PropertyGroup>
 

src/TopDownProteomics/ProForma/ProFormaParser.cs

@@ -10,10 +10,26 @@ namespace TopDownProteomics.ProForma
     /// </summary>
     public class ProFormaParser
     {
+        private readonly Func<char, bool> _residueValidator;
+
         /// <summary>
         /// Initializes a new instance of the <see cref="ProFormaParser"/> class.
         /// </summary>
-        public ProFormaParser() { }
+        public ProFormaParser() : this(char.IsUpper) { }
+
+        /// <summary>
+        /// Initializes a new instance of the <see cref="ProFormaParser"/> class with a residue validator.
+        /// </summary>
+        /// <param name="residueValidator">Determines whether a character is a valid residue.</param>
+        /// <exception cref="ArgumentNullException"><paramref name="residueValidator"/> is <see langword="null"/>.</exception>
+        /// <remarks>
+        /// For workflows that create a proteoform group or chemical hash, use a matching
+        /// <see cref="TopDownProteomics.Biochemistry.IResidueProvider"/> with the custom residue validator.
+        /// </remarks>
+        public ProFormaParser(Func<char, bool> residueValidator)
+        {
+            _residueValidator = residueValidator ?? throw new ArgumentNullException(nameof(residueValidator));
+        }
 
 #if !NETSTANDARD2_1
         /// <summary>
@@ -218,8 +234,8 @@ public ProFormaTerm ParseString(ReadOnlySpan<char> proFormaString)
                 else
                 {
                     // Validate amino acid character
-                    if (!char.IsUpper(current))
-                        throw new ProFormaParseException($"{current} is not an upper case letter.");
+                    if (!_residueValidator(current))
+                        throw new ProFormaParseException($"{current} is not a valid residue.");
 
                     // Reset the range as soon as we see an amino acid
                     if (endRange.HasValue)
@@ -263,9 +279,12 @@ private void HandleGlobalModification(ref IDictionary<string, ProFormaTagGroup>?
                 targets = new List<char>();
                 for (int k = atSymbolIndex + 1; k < tagText.Length; k++)
                 {
-                    if (char.IsUpper(tagText[k]))
+                    if (tagText[k] == ',')
+                        continue;
+
+                    if (_residueValidator(tagText[k]))
                         targets.Add(tagText[k]);
-                    else if (tagText[k] != ',')
+                    else
                         throw new ProFormaParseException($"Unexpected character {tagText[k]} in global modification target list.");
                 }
             }

tests/TopDownProteomics.Tests/ProForma/ProFormaParserTests.cs

@@ -15,6 +15,7 @@ public void InvalidProFormaStrings()
         {
             Assert.Throws<ArgumentNullException>(() => _parser.ParseString((Span<char>)null));
             Assert.Throws<ArgumentNullException>(() => _parser.ParseString(string.Empty));
+            Assert.Throws<ArgumentNullException>(() => new ProFormaParser(null!));
         }
 
         [Test]
@@ -27,6 +28,24 @@ public void NoModifications()
             Assert.IsNull(term.Tags);
         }
 
+        [Test]
+        public void ResidueValidatorControlsSequenceAndGlobalModificationTargets()
+        {
+            var parser = new ProFormaParser(residue => residue == 'A' || residue == '*');
+
+            var term = parser.ParseString("<[Oxidation]@*,A>A*");
+
+            Assert.AreEqual("A*", term.Sequence);
+            CollectionAssert.AreEquivalent(new[] { '*', 'A' }, term.GlobalModifications.Single().TargetAminoAcids);
+            var exception = Assert.Throws<ProFormaParseException>(() => parser.ParseString("AB"));
+            Assert.AreEqual("B is not a valid residue.", exception.Message);
+            Assert.Throws<ProFormaParseException>(() => _parser.ParseString("A*"));
+
+            var permissiveParser = new ProFormaParser(_ => true);
+            term = permissiveParser.ParseString("<[Oxidation]@A,C>A");
+            CollectionAssert.AreEquivalent(new[] { 'A', 'C' }, term.GlobalModifications.Single().TargetAminoAcids);
+        }
+
         [Test]
         [TestCase("PRO[info:test]TEOFORM", "PROTEOFORM", "test")]
         [TestCase("PRO[info:test[nested]]TEOFORM", "PROTEOFORM", "test[nested]")]

tests/TopDownProteomics.Tests/TopDownProteomics.Tests.csproj

@@ -1,6 +1,6 @@
 <Project Sdk="Microsoft.NET.Sdk">
   <PropertyGroup>
-    <TargetFrameworks>net6.0;net472</TargetFrameworks>
+    <TargetFrameworks>net8.0;net472</TargetFrameworks>
     <RootNamespace>TopDownProteomics.Tests</RootNamespace>
     <Authors>Consortium for Top-Down Proteomics</Authors>
     <LangVersion>latest</LangVersion>