Fix input to allow empty fixed modifications from SDRF

CHANGELOG.md

@@ -3,6 +3,13 @@
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/)
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [Unreleased]
+
+### `Fixed`
+
+- Allow an empty fixed-modification set from SDRF without adding a default modification. Reject searches using MS-GF+ when both modification sets remain empty after the variable-modification fallback, to prevent OpenMS MSGFPlusAdapter from silently enabling fixed Carbamidomethyl (C).
+- Trim the selected variable-modification value before validation so a whitespace-only fallback cannot bypass the MS-GF+ empty-modification check.
+
 ## [1.10.0] bigbio/quantms - [15/08/2026] - [London]
 
 ### `Changed`

docs/usage.md

@@ -16,6 +16,14 @@ nextflow run bigbio/quantms --input '/url/path/to/your/experiment_design.sdrf.ts
 
 The input file must be in [Sample-to-data-relationship format (SDRF)](https://pubs.acs.org/doi/abs/10.1021/acs.jproteome.0c00376) and can have `.sdrf`, `.tsv`, or `.csv` file extensions.
 
+### Search modifications from SDRF
+
+Fixed modifications are read exclusively from the SDRF modification annotations. If the parsed `FixedModifications` value is empty, quantms preserves an empty fixed-modification set instead of rejecting the input or adding a default modification such as Carbamidomethyl (C). Annotate modifications according to the experimental protocol; a modification that is variable must not be declared fixed just to make the input pass validation.
+
+Variable modifications are read from the SDRF when present. If the parsed variable-modification set is empty or contains only whitespace, quantms uses `--variable_mods`, whose default is `Oxidation (M)`. Leading and trailing whitespace is removed from the selected SDRF or fallback value before validation; a whitespace-only fallback is treated as an empty set. An empty fixed-modification set with a nonempty variable-modification set is supported with Comet, MS-GF+, and Sage.
+
+For a search without either fixed or variable modifications, select Comet and/or Sage and set `variable_mods` to an empty string. When `--search_engines` includes `msgf`, quantms rejects an empty final set of both modification types: the current OpenMS MSGFPlusAdapter would otherwise enable fixed Carbamidomethyl (C). This check runs after the variable-modification fallback has been applied.
+
 ### Supported file formats
 
 The pipeline supports the following mass spectrometry data file formats:

nextflow_schema.json

@@ -19,7 +19,7 @@
                     "mimetype": "text/csv",
                     "pattern": "^\\S+\\.(?:csv|tsv|sdrf)$",
                     "description": "URI/path to an SDRF file in SDRF format with .sdrf, .tsv, or .csv extension. For more info see help text or docs.",
-                    "help_text": "Input is specified by using a path or URI to a PRIDE Sample to Data Relation Format file (SDRF), e.g. as part of a submitted and annotated PRIDE experiment (see [here](https://github.com/bigbio/proteomics-metadata-standard/tree/master/annotated-projects) for examples). Input files will be downloaded and cached from the URIs specified in the SDRF file.\n\nThe SDRF file can have .sdrf, .tsv, or .csv extensions. An OpenMS-style experimental design will be generated based on the factor columns of the SDRF.\n\nThe following parameters are read **exclusively** from the SDRF file (required columns):\n\n    * `acquisition_method` (from 'Proteomics Data Acquisition Method' column),\n    * `labelling_type` (from 'Label' column),\n    * `enzyme` (from 'Enzyme' column),\n    * `fixed_mods` (from 'FixedModifications' column)\n\nThe following parameters are read from SDRF but can be **overridden via command line**:\n\n    * `precursor_mass_tolerance` (from 'PrecursorMassTolerance' column),\n    * `precursor_mass_tolerance_unit` (from 'PrecursorMassToleranceUnit' column),\n    * `fragment_mass_tolerance` (from 'FragmentMassTolerance' column),\n    * `fragment_mass_tolerance_unit` (from 'FragmentMassToleranceUnit' column),\n    * `variable_mods` (from 'VariableModifications' column)",
+                    "help_text": "Input is specified by using a path or URI to a PRIDE Sample to Data Relation Format file (SDRF), e.g. as part of a submitted and annotated PRIDE experiment (see [here](https://github.com/bigbio/proteomics-metadata-standard/tree/master/annotated-projects) for examples). Input files will be downloaded and cached from the URIs specified in the SDRF file.\n\nThe SDRF file can have .sdrf, .tsv, or .csv extensions. An OpenMS-style experimental design will be generated based on the factor columns of the SDRF.\n\nThe following parameters are read **exclusively** from the SDRF file (required columns):\n\n    * `acquisition_method` (from 'Proteomics Data Acquisition Method' column),\n    * `labelling_type` (from 'Label' column),\n    * `enzyme` (from 'Enzyme' column)\n\nFixed modifications are also read exclusively from the SDRF ('FixedModifications' column). An empty set is supported and does not add a default fixed modification. When using MS-GF+, at least one fixed or variable modification must remain after applying the variable-modification fallback, because OpenMS MSGFPlusAdapter otherwise enables fixed Carbamidomethyl (C). Use Comet or Sage for a search without modifications.\n\nThe following parameters are read from SDRF but can be **overridden via command line**:\n\n    * `precursor_mass_tolerance` (from 'PrecursorMassTolerance' column),\n    * `precursor_mass_tolerance_unit` (from 'PrecursorMassToleranceUnit' column),\n    * `fragment_mass_tolerance` (from 'FragmentMassTolerance' column),\n    * `fragment_mass_tolerance_unit` (from 'FragmentMassToleranceUnit' column),\n    * `variable_mods` (from 'VariableModifications' column)",
                     "fa_icon": "fas fa-file-csv"
                 },
                 "outdir": {

subworkflows/local/create_input_channel/main.nf

@@ -105,8 +105,7 @@ def create_meta_channel(LinkedHashMap row, enzymes, files, wrapper) {
     // Validate required SDRF columns - these parameters are exclusively read from SDRF (no command-line override)
     def requiredColumns = [
         'Label': row.Label,
-        'Enzyme': row.Enzyme,
-        'FixedModifications': row.FixedModifications
+        'Enzyme': row.Enzyme
     ]
 
     def missingColumns = []
@@ -122,9 +121,9 @@ def create_meta_channel(LinkedHashMap row, enzymes, files, wrapper) {
         exit(1)
     }
 
-    // Set values from SDRF (required columns)
+    // An empty fixed-modification set must remain an empty string for the search adapters.
     meta.labelling_type = row.Label
-    meta.fixedmodifications = row.FixedModifications
+    meta.fixedmodifications = row.FixedModifications?.toString()?.trim() ?: ''
     meta.enzyme = row.Enzyme
 
     // Set tolerance values: use SDRF if available, otherwise fall back to params
@@ -182,6 +181,13 @@ def create_meta_channel(LinkedHashMap row, enzymes, files, wrapper) {
     } else {
         meta.variablemodifications = params.variable_mods
     }
+    meta.variablemodifications = meta.variablemodifications?.toString()?.trim() ?: ''
+
+    if (params.search_engines.contains('msgf') && !meta.fixedmodifications && !meta.variablemodifications) {
+        exit(1, "ERROR: Both modification lists are empty for '${filestr}', " +
+            'but OpenMS MSGFPlusAdapter would enable fixed Carbamidomethyl (C) in this case. ' +
+            'Use Comet or Sage for a search without modifications.')
+    }
 
     enzymes += row.Enzyme
     if (enzymes.size() > 1) {

subworkflows/local/create_input_channel/tests/main.nf.test

@@ -0,0 +1,97 @@
+nextflow_function {
+
+    name "SDRF search modification metadata"
+    script "../main.nf"
+    function "create_meta_channel"
+    tag "create_input_channel"
+
+    [
+        [name: 'empty fixed set', fixed: '', variable: 'Carbamidomethyl (C),Oxidation (M),Propionamide (C)'],
+        [name: 'missing fixed set', fixed: null, variable: 'Oxidation (M)'],
+        [
+            name: 'whitespace fixed set', fixed: '  ',
+            variable: 'Carbamidomethyl (C),Oxidation (M),Acetyl (Protein N-term)'
+        ],
+        [name: 'nonempty fixed set', fixed: 'Carbamidomethyl (C)', variable: 'Oxidation (M)'],
+        [name: 'variable fallback', fixed: '', variable: null],
+        [
+            name: 'trim SDRF variable set', fixed: '', variable: '  Oxidation (M)  ',
+            fallback: 'Acetyl (Protein N-term)', expectedVariable: 'Oxidation (M)'
+        ],
+        [
+            name: 'trim variable fallback', fixed: '', variable: null,
+            fallback: '  Oxidation (M)  ', expectedVariable: 'Oxidation (M)'
+        ],
+        [
+            name: 'whitespace SDRF uses fallback', fixed: '', variable: '  ',
+            expectedVariable: 'Oxidation (M)'
+        ],
+        [name: 'no modifications with Comet', fixed: '', variable: '', fallback: '', engines: 'comet'],
+        [name: 'no modifications with Sage', fixed: '', variable: '', fallback: '', engines: 'sage'],
+        [
+            name: 'whitespace fallback with Comet', fixed: '', variable: '', fallback: '  ',
+            engines: 'comet', expectedVariable: ''
+        ],
+        [
+            name: 'null fallback with Sage', fixed: '', variable: null, fallback: null,
+            engines: 'sage', expectedVariable: ''
+        ],
+        [
+            name: 'reject MSGF implicit CAM', fixed: '', variable: '', fallback: '',
+            error: 'Both modification lists are empty'
+        ],
+        [
+            name: 'reject MSGF whitespace fallback', fixed: '', variable: '', fallback: '  ',
+            error: 'Both modification lists are empty'
+        ],
+        [
+            name: 'require label', fixed: '', variable: 'Oxidation (M)', label: '',
+            error: 'Missing or empty required SDRF columns'
+        ],
+        [
+            name: 'require enzyme', fixed: '', variable: 'Oxidation (M)', enzyme: '',
+            error: 'Missing or empty required SDRF columns'
+        ],
+    ].each { example ->
+        test(example.name) {
+            when {
+                params {
+                    root_folder = null
+                    search_engines = example.engines ?: 'comet,msgf,sage'
+                    variable_mods = example.containsKey('fallback') ? example.fallback : 'Oxidation (M)'
+                }
+                function {
+                    """
+                    input[0] = [
+                        URI: "${projectDir}/main.nf",
+                        'Proteomics Data Acquisition Method': 'data-dependent acquisition',
+                        DissociationMethod: 'BEAM-TYPE COLLISION-INDUCED DISSOCIATION',
+                        Label: ${(example.containsKey('label') ? example.label : 'label free sample').inspect()},
+                        Enzyme: ${(example.containsKey('enzyme') ? example.enzyme : 'Trypsin').inspect()},
+                        FixedModifications: ${example.fixed.inspect()},
+                        VariableModifications: ${example.variable.inspect()}
+                    ]
+                    input[1] = [] as Set
+                    input[2] = [] as Set
+                    input[3] = [labelling_type: '', experiment_id: 'test.sdrf.tsv']
+                    """
+                }
+            }
+            then {
+                if (example.error) {
+                    assert function.failed
+                    assert function.stdout.any { it.contains(example.error) }
+                } else {
+                    assert function.success
+                    assert function.result[0].fixedmodifications == (example.fixed?.trim() ?: '')
+                    def fallback = example.containsKey('fallback') ? example.fallback : 'Oxidation (M)'
+                    def expectedVariable = example.containsKey('expectedVariable')
+                        ? example.expectedVariable : (example.variable ?: fallback)
+                    assert function.result[0].variablemodifications == expectedVariable
+                    assert function.result[0].enzyme == 'Trypsin'
+                    assert function.result[0].labelling_type == 'label free sample'
+                }
+            }
+        }
+    }
+}