From 3ca2297c63917a022081d13062030d34ae2b4fb9 Mon Sep 17 00:00:00 2001
From: Will Fondrie <wfondrie@talus.bio>
Date: Tue, 7 May 2024 22:59:38 -0700
Subject: [PATCH] Add support for unsigned masses in otherwise valid Proforma
 peptides. (#52)

* Add support for sign missing in proforma string

* Minor refactoring and update

* Bump depenency versions
---
 CHANGELOG.md                                  |  4 ++
 depthcharge/primitives.py                     | 42 +++++++++++++------
 pyproject.toml                                |  3 +-
 tests/unit_tests/test_primitives.py           |  6 +++
 .../test_tokenizers/test_peptides.py          |  6 +++
 5 files changed, 48 insertions(+), 13 deletions(-)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index 8be1135..e49edc2 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -6,6 +6,10 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [v0.4.6]
+### Added
+- Added support for unsigned modification masses that don't quite conform to the Proforma standard.
+
 ## [v0.4.5]
 ### Changed
 - The `scan_id` column for parsed spectra is not a sting instead of an integer. This is less space efficient, but we ran into issues with Sciex indexing when trying to use only an integer.
diff --git a/depthcharge/primitives.py b/depthcharge/primitives.py
index b7b3e3a..d81382b 100644
--- a/depthcharge/primitives.py
+++ b/depthcharge/primitives.py
@@ -17,6 +17,7 @@
 from rdkit.Chem import Draw
 from spectrum_utils.spectrum import MsmsSpectrum
 
+from . import utils
 from .constants import PROTON
 
 MSKB_TO_UNIMOD = {
@@ -72,17 +73,7 @@ def __post_init__(self) -> None:
             if mod is None:
                 continue
 
-            try:
-                mod = [MassModification(mod)]
-            except ValueError:
-                try:
-                    mod = [GenericModification(mod)]
-                except (AttributeError, TypeError):
-                    pass
-            except TypeError:
-                pass
-
-            parsed[idx] = mod
+            parsed[idx] = [_resolve_mod(m) for m in utils.listify(mod)]
 
         self.modifications = parsed
         n_mod = self.modifications[0]
@@ -116,7 +107,7 @@ def split(self) -> list[str]:
                 except (AttributeError, ValueError):
                     modstr = f"[{mods[0].mass:+0.6f}]"
             else:
-                modstr = f"[{sum([m.mass for m in mods]):+0.6f}]"
+                modstr = f"[{sum(m.mass for m in mods):+0.6f}]"
 
             if not idx:
                 out.append(f"{modstr}-")
@@ -444,3 +435,30 @@ def to_tensor(self) -> torch.tensor:
 
         """
         return torch.tensor(np.vstack([self.mz, self.intensity]).T)
+
+
+def _resolve_mod(
+    mod: MassModification | GenericModification | str | float,
+) -> MassModification | GenericModification:
+    """Resolve the type of a modification.
+
+    Parameters
+    ----------
+    mod : MassModification, GenericModification, str, or float
+        The modification to resolve.
+
+    Returns
+    -------
+    MassModification or GenericModification
+        The best modification for the input type.
+
+    """
+    try:
+        mod = mod.value
+    except AttributeError:
+        pass
+
+    try:
+        return MassModification(float(mod))
+    except ValueError:
+        return GenericModification(str(mod))
diff --git a/pyproject.toml b/pyproject.toml
index 8d41c1b..8638a95 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -19,7 +19,7 @@ dependencies = [
     "polars>=0.19.0",
     "pyarrow>=12.0.1",
     "pylance>=0.7.5",
-    "pyteomics>=4.4.2",
+    "pyteomics>=4.7.2",
     "numpy>=1.18.1",
     "numba>=0.48.0",
     "lxml>=4.9.1",
@@ -33,6 +33,7 @@ dependencies = [
     "pillow>=9.4.0",
     "spectrum-utils>=0.4.1",
     "cloudpathlib>=0.18.1",
+    "psims>=1.3.3"
 ]
 dynamic = ["version"]
 
diff --git a/tests/unit_tests/test_primitives.py b/tests/unit_tests/test_primitives.py
index 19fb2bc..631d4cd 100644
--- a/tests/unit_tests/test_primitives.py
+++ b/tests/unit_tests/test_primitives.py
@@ -50,6 +50,12 @@ def test_peptide_init():
     assert parsed.split() == expected
 
 
+def test_almost_proforma():
+    """Test a peptide lacking an explicit sign."""
+    parsed = Peptide.from_proforma("LES[79.0]LIEK")
+    assert parsed.split() == ["L", "E", "S[+79.000000]", "L", "I", "E", "K"]
+
+
 def test_peptide_from_proforma():
     """Test proforma parsing."""
     parsed = Peptide.from_proforma("LESLIEK/2")
diff --git a/tests/unit_tests/test_tokenizers/test_peptides.py b/tests/unit_tests/test_tokenizers/test_peptides.py
index bfd8785..b5f58e9 100644
--- a/tests/unit_tests/test_tokenizers/test_peptides.py
+++ b/tests/unit_tests/test_tokenizers/test_peptides.py
@@ -109,3 +109,9 @@ def test_single_peptide():
     ion = tokenizer.calculate_precursor_ions("LESLIEK", 2)
     expected = mass.fast_mass("LESLIEK", charge=2, ion_type="M")
     torch.testing.assert_close(ion, torch.tensor([expected]))
+
+
+def test_almost_compliant_proform():
+    """Test initializing with a peptide without an expicit mass sign."""
+    tokenizer = PeptideTokenizer.from_proforma("[10]-EDITHR")
+    assert "[+10.000000]-" in tokenizer.residues