PM4 and BP3 rules for VCEPs (#177)

* PM4 and BP3 rules for VCEPs

* Unit tests

* fix integ test for PM4

* PTEN unit test

* formatting

.vscode/launch.json

@@ -6,7 +6,7 @@
           "type": "debugpy",
           "request": "launch",
           "module": "src.cli",
-          "args": ["NM_000038.6(APC):c.2476T>G", "--genome-release", "grch38"],
+          "args": ["NM_000314.7(PTEN):c.1133_1136del", "--genome-release", "grch38"],
           "console": "integratedTerminal"
       },
       {

src/criteria/auto_pm4_bp3.py

@@ -99,6 +99,20 @@ def is_inframe_delins(var_data: AutoACMGData) -> bool:
             return True
         return False
 
+    def _bp3_not_applicable(self, seqvar: SeqVar, var_data: AutoACMGData) -> bool:
+        """
+        Check if BP3 is not applicable for the variant.
+
+        Args:
+            var_data (AutoACMGData): The variant information.
+
+        Returns:
+            bool: True if BP3 is not applicable, False otherwise.
+        """
+        if seqvar.chrom == "MT":
+            return True
+        return False
+
     def verify_pm4bp3(self, seqvar: SeqVar, var_data: AutoACMGData) -> Tuple[Optional[PM4BP3], str]:
         """Predicts PM4 and BP3 criteria for the provided sequence variant.
 
@@ -182,6 +196,23 @@ def predict_pm4bp3(
             bp3_pred = AutoACMGPrediction.Failed
             pm4_strength = AutoACMGStrength.PathogenicModerate
             bp3_strength = AutoACMGStrength.BenignSupporting
+
+        # BP3 is not applicable for some VCEPs
+        if self._bp3_not_applicable(seqvar, var_data):
+            return (
+                AutoACMGCriteria(
+                    name="PM4",
+                    prediction=pm4_pred,
+                    strength=pm4_strength,
+                    summary=comment,
+                ),
+                AutoACMGCriteria(
+                    name="BP3",
+                    prediction=AutoACMGPrediction.NotApplicable,
+                    strength=bp3_strength,
+                    summary="BP3 is not applicable for the gene.",
+                ),
+            )
         return (
             AutoACMGCriteria(
                 name="PM4",

src/defs/auto_acmg.py

@@ -1,6 +1,8 @@
 from enum import auto
 from typing import Dict, List, Optional
 
+from pydantic import BaseModel, ConfigDict
+
 from src.defs.annonars_variant import GnomadExomes, GnomadMtDna
 from src.defs.core import AutoAcmgBaseEnum, AutoAcmgBaseModel
 from src.defs.mehari import Exon, TranscriptGene, TranscriptSeqvar
@@ -428,3 +430,17 @@ class AutoACMGResult(AutoAcmgBaseModel):
     data: AutoACMGData = AutoACMGData()
     # ; ACMG criteria prediction
     criteria: AutoACMGCriteriaResult = AutoACMGCriteriaResult()
+
+
+class VcepSpec(BaseModel):
+    #: Identifier, e.g., "GN002"
+    identifier: str
+    #: Version, e.g., "2.0.0"
+    version: str
+    #: Title of the VCEP specification, e.g., "ClinGen Cardiomyopathy Expert Panel Specifications
+    #: to the ACMG/AMP Variant Interpretation Guidelines for MYH7".
+    title: Optional[str] = None
+
+    model_config = ConfigDict(
+        frozen=True,
+    )

src/vcep/__init__.py

@@ -15,7 +15,7 @@
 from src.vcep.glaucoma import GlaucomaPredictor
 from src.vcep.hbopc import HBOPCPredictor
 from src.vcep.hearing_loss import HearingLossPredictor
-from src.vcep.hhtp import HHTPredictor
+from src.vcep.hht import HHTPredictor
 from src.vcep.insight_colorectal_cancer import InsightColorectalCancerPredictor
 from src.vcep.leber_congenital_amaurosis import LeberCongenitalAmaurosisPredictor
 from src.vcep.lysosomal_diseases import LysosomalDiseasesPredictor

src/vcep/acadvl.py

@@ -4,12 +4,26 @@
 Link: https://cspec.genome.network/cspec/ui/svi/doc/GN021
 """
 
+from typing import Tuple
+
 from loguru import logger
 
 from src.criteria.default_predictor import DefaultPredictor
-from src.defs.auto_acmg import AutoACMGCriteria, AutoACMGData, AutoACMGPrediction, AutoACMGStrength
+from src.defs.auto_acmg import (
+    AutoACMGCriteria,
+    AutoACMGData,
+    AutoACMGPrediction,
+    AutoACMGStrength,
+    VcepSpec,
+)
 from src.defs.seqvar import SeqVar
 
+#: VCEP specification for ACADVL.
+SPEC: VcepSpec = VcepSpec(
+    identifier="GN021",
+    version="1.0.0",
+)
+
 PM1_CLUSTER = [
     (214, 223),  # Nucleotide and substrate binding
     (249, 251),  # Nucleotide and substrate binding
@@ -44,3 +58,7 @@ def predict_pm1(self, seqvar: SeqVar, var_data: AutoACMGData) -> AutoACMGCriteri
             strength=AutoACMGStrength.PathogenicModerate,
             summary="Variant does not fall within any critical region for ACADVL. PM1 is not met.",
         )
+
+    def _bp3_not_applicable(self, seqvar: SeqVar, var_data: AutoACMGData) -> bool:
+        """Override BP3 to be not applicable for ACADVL."""
+        return True

src/vcep/brain_malformations.py

@@ -74,6 +74,26 @@ def predict_pm1(self, seqvar: SeqVar, var_data: AutoACMGData) -> AutoACMGCriteri
         else:
             return super().predict_pm1(seqvar, var_data)
 
+    def predict_pm4bp3(
+        self, seqvar: SeqVar, var_data: AutoACMGData
+    ) -> Tuple[AutoACMGCriteria, AutoACMGCriteria]:
+        """Override predict_pm4bp3 to include VCEP-specific logic for brain malformations VCEP."""
+        logger.info("Predict PM4 and BP3")
+        return (
+            AutoACMGCriteria(
+                name="PM4",
+                prediction=AutoACMGPrediction.NotApplicable,
+                strength=AutoACMGStrength.PathogenicModerate,
+                summary="PM4 is not applicable for the brain malformations VCEP.",
+            ),
+            AutoACMGCriteria(
+                name="BP3",
+                prediction=AutoACMGPrediction.NotApplicable,
+                strength=AutoACMGStrength.BenignSupporting,
+                summary="BP3 is not applicable for the brain malformations VCEP.",
+            ),
+        )
+
     def predict_bp7(self, seqvar: SeqVar, var_data: AutoACMGData) -> AutoACMGCriteria:
         """Change the PhyloP100 score threshold for BP7."""
         var_data.thresholds.phyloP100 = 0.1

src/vcep/cardiomyopathy.py

@@ -20,6 +20,8 @@
 https://cspec.genome.network/cspec/ui/svi/doc/GN103
 """
 
+from typing import Tuple
+
 from loguru import logger
 
 from src.criteria.default_predictor import DefaultPredictor
@@ -99,6 +101,10 @@ def predict_pm1(self, seqvar: SeqVar, var_data: AutoACMGData) -> AutoACMGCriteri
 
         return super().predict_pm1(seqvar, var_data)
 
+    def _bp3_not_applicable(self, seqvar: SeqVar, var_data: AutoACMGData) -> bool:
+        """Override BP3 for Cardiomyopathy."""
+        return True
+
     def predict_bp7(self, seqvar: SeqVar, var_data: AutoACMGData) -> AutoACMGCriteria:
         """Override donor and acceptor positions for Cardiomyopathy VCEP."""
         var_data.thresholds.bp7_donor = 7

src/vcep/cdh1.py

@@ -4,6 +4,8 @@
 Link: https://cspec.genome.network/cspec/ui/svi/doc/GN007
 """
 
+from typing import Tuple
+
 from loguru import logger
 
 from src.criteria.default_predictor import DefaultPredictor
@@ -24,6 +26,36 @@ def predict_pm1(self, seqvar: SeqVar, var_data: AutoACMGData) -> AutoACMGCriteri
             summary="PM1 is not applicable for CDH1.",
         )
 
+    def predict_pm4bp3(
+        self, seqvar: SeqVar, var_data: AutoACMGData
+    ) -> Tuple[AutoACMGCriteria, AutoACMGCriteria]:
+        """Override predict_pm4bp3 to include VCEP-specific logic for CDH1."""
+        logger.info("Predict PM4 and BP3")
+        pred, comment = super().verify_pm4bp3(seqvar, var_data)
+        if pred:
+            pm4 = (
+                AutoACMGPrediction.Met
+                if pred.PM4
+                else (AutoACMGPrediction.NotMet if pred.PM4 is False else AutoACMGPrediction.Failed)
+            )
+        else:
+            pm4 = AutoACMGPrediction.Failed
+            comment = "PM4 could not be verified."
+        return (
+            AutoACMGCriteria(
+                name="PM4",
+                prediction=pm4,
+                strength=AutoACMGStrength.PathogenicSupporting,
+                summary=comment,
+            ),
+            AutoACMGCriteria(
+                name="BP3",
+                prediction=AutoACMGPrediction.NotApplicable,
+                strength=AutoACMGStrength.BenignSupporting,
+                summary="BP3 is not applicable for CDH1.",
+            ),
+        )
+
     def predict_bp7(self, seqvar: SeqVar, var_data: AutoACMGData) -> AutoACMGCriteria:
         """Override donor and acceptor positions for CDH1 VCEP."""
         var_data.thresholds.bp7_donor = 7

src/vcep/cerebral_creatine_deficiency_syndromes.py

@@ -10,6 +10,8 @@
 https://cspec.genome.network/cspec/ui/svi/doc/GN027
 """
 
+from typing import Tuple
+
 from src.criteria.default_predictor import DefaultPredictor
 from src.defs.auto_acmg import AutoACMGCriteria, AutoACMGData, AutoACMGPrediction, AutoACMGStrength
 from src.defs.seqvar import SeqVar
@@ -29,6 +31,10 @@ def predict_pm1(self, seqvar: SeqVar, var_data: AutoACMGData) -> AutoACMGCriteri
             summary="PM1 is not applicable for Cerebral Creatine Deficiency Syndromes.",
         )
 
+    def _bp3_not_applicable(self, seqvar: SeqVar, var_data: AutoACMGData) -> bool:
+        """Override _bp3_not_applicable for Cerebral Creatine Deficiency Syndromes."""
+        return True
+
     def predict_bp7(self, seqvar: SeqVar, var_data: AutoACMGData) -> AutoACMGCriteria:
         """Override donor and acceptor positions for Cerebral Creatine Deficiency Syndromes VCEP."""
         if var_data.hgnc_id == "HGNC:4136":

src/vcep/coagulation_factor_deficiency.py

@@ -8,7 +8,7 @@
 https://cspec.genome.network/cspec/ui/svi/doc/GN080
 """
 
-from typing import Dict, List
+from typing import Dict, List, Tuple
 
 from loguru import logger
 
@@ -133,6 +133,10 @@ def predict_pm1(self, seqvar: SeqVar, var_data: AutoACMGData) -> AutoACMGCriteri
             summary=f"Variant does not meet the PM1 criteria for {var_data.hgnc_id}.",
         )
 
+    def _bp3_not_applicable(self, seqvar: SeqVar, var_data: AutoACMGData) -> bool:
+        """Override BP3 for Coagulation Factor Deficiency."""
+        return True
+
     def predict_bp7(self, seqvar: SeqVar, var_data: AutoACMGData) -> AutoACMGCriteria:
         """Change the spliceAI and phyloP threshold for BP7."""
         if var_data.hgnc_id == "HGNC:3546":

src/vcep/congenital_myopathies.py

@@ -63,3 +63,7 @@ def predict_pm1(self, seqvar: SeqVar, var_data: AutoACMGData) -> AutoACMGCriteri
             strength=AutoACMGStrength.PathogenicModerate,
             summary="Variant does not fall within a critical domain.",
         )
+
+    def _bp3_not_applicable(self, seqvar: SeqVar, var_data: AutoACMGData) -> bool:
+        """Override BP3 for congenital myopathies."""
+        return True

src/vcep/dicer1.py

@@ -76,6 +76,10 @@ def predict_pm1(self, seqvar: SeqVar, var_data: AutoACMGData) -> AutoACMGCriteri
             summary="Variant does not affect a critical domain for DICER1.",
         )
 
+    def _bp3_not_applicable(self, seqvar: SeqVar, var_data: AutoACMGData) -> bool:
+        """Override BP3 to be not applicable for DICER1."""
+        return True
+
     def predict_bp7(self, seqvar: SeqVar, var_data: AutoACMGData) -> AutoACMGCriteria:
         """Override donor and acceptor positions for DICER1 VCEP."""
         var_data.thresholds.bp7_donor = 7

src/vcep/enigma.py

@@ -52,6 +52,26 @@ def predict_pm1(self, seqvar: SeqVar, var_data: AutoACMGData) -> AutoACMGCriteri
 
         return super().predict_pm1(seqvar, var_data)
 
+    def predict_pm4bp3(
+        self, seqvar: SeqVar, var_data: AutoACMGData
+    ) -> Tuple[AutoACMGCriteria, AutoACMGCriteria]:
+        """Override predict_pm4bp3 to include VCEP-specific logic for ENIGMA BRCA1 and BRCA2."""
+        logger.info("Predict PM4 and BP3")
+        return (
+            AutoACMGCriteria(
+                name="PM4",
+                prediction=AutoACMGPrediction.NotApplicable,
+                strength=AutoACMGStrength.PathogenicModerate,
+                summary="PM4 is not applicable for the ENIGMA VCEP.",
+            ),
+            AutoACMGCriteria(
+                name="BP3",
+                prediction=AutoACMGPrediction.NotApplicable,
+                strength=AutoACMGStrength.BenignSupporting,
+                summary="BP3 is not applicable for the ENIGMA VCEP.",
+            ),
+        )
+
     def _in_important_domain(self, var_data: AutoACMGData) -> bool:
         """Check if the variant is in an important domain."""
         for start, end in BP7_IMPORTANT_DOMAINS.get(var_data.hgnc_id, []):

src/vcep/familial_hypercholesterolemia.py

@@ -67,3 +67,7 @@ def predict_pm1(self, seqvar: SeqVar, var_data: AutoACMGData) -> AutoACMGCriteri
             strength=AutoACMGStrength.PathogenicModerate,
             summary=f"Variant does not meet the PM1 criteria for LDLR.",
         )
+
+    def _bp3_not_applicable(self, seqvar: SeqVar, var_data: AutoACMGData) -> bool:
+        """Override BP3 for Familial Hypercholesterolemia."""
+        return True

src/vcep/fbn1.py

@@ -88,3 +88,7 @@ def predict_pm1(self, seqvar: SeqVar, var_data: AutoACMGData) -> AutoACMGCriteri
             strength=AutoACMGStrength.PathogenicModerate,
             summary=f"Variant does not meet the PM1 criteria for FBN1.",
         )
+
+    def _bp3_not_applicable(self, seqvar: SeqVar, var_data: AutoACMGData) -> bool:
+        """Override BP3 for FBN1."""
+        return True

src/vcep/glaucoma.py

@@ -25,6 +25,10 @@ def predict_pm1(self, seqvar: SeqVar, var_data: AutoACMGData) -> AutoACMGCriteri
             summary="PM1 is not applicable for MYOC.",
         )
 
+    def _bp3_not_applicable(self, seqvar: SeqVar, var_data: AutoACMGData) -> bool:
+        """Override _bp3_not_applicable for Glaucoma VCEP."""
+        return True
+
     def _is_conserved(self, var_data: AutoACMGData) -> bool:
         """
         Predict if the variant is conserved.

src/vcep/hbopc.py

@@ -8,6 +8,8 @@
 https://cspec.genome.network/cspec/ui/svi/doc/GN077
 """
 
+from typing import Tuple
+
 from loguru import logger
 
 from src.criteria.default_predictor import DefaultPredictor
@@ -37,6 +39,58 @@ def predict_pm1(self, seqvar: SeqVar, var_data: AutoACMGData) -> AutoACMGCriteri
 
         return super().predict_pm1(seqvar, var_data)
 
+    def predict_pm4bp3(
+        self, seqvar: SeqVar, var_data: AutoACMGData
+    ) -> Tuple[AutoACMGCriteria, AutoACMGCriteria]:
+        """Override predict_pm4bp3 to include VCEP-specific logic for CDH1."""
+        logger.info("Predict PM4 and BP3")
+
+        if var_data.hgnc_id == "HGNC:795":
+            pred, comment = self.verify_pm4bp3(seqvar, var_data)
+            if pred:
+                pm4 = (
+                    AutoACMGPrediction.Met
+                    if pred.PM4
+                    else (
+                        AutoACMGPrediction.NotMet
+                        if pred.PM4 is False
+                        else AutoACMGPrediction.Failed
+                    )
+                )
+            else:
+                pm4 = AutoACMGPrediction.Failed
+                comment = "PM4 could not be verified."
+            return (
+                AutoACMGCriteria(
+                    name="PM4",
+                    prediction=pm4,
+                    strength=AutoACMGStrength.PathogenicSupporting,
+                    summary=comment,
+                ),
+                AutoACMGCriteria(
+                    name="BP3",
+                    prediction=AutoACMGPrediction.NotApplicable,
+                    strength=AutoACMGStrength.BenignSupporting,
+                    summary="BP3 is not applicable for ATM.",
+                ),
+            )
+        elif var_data.hgnc_id == "HGNC:26144":
+            return (
+                AutoACMGCriteria(
+                    name="PM4",
+                    prediction=AutoACMGPrediction.NotApplicable,
+                    strength=AutoACMGStrength.PathogenicSupporting,
+                    summary="PM4 is not applicable for PALB2.",
+                ),
+                AutoACMGCriteria(
+                    name="BP3",
+                    prediction=AutoACMGPrediction.NotApplicable,
+                    strength=AutoACMGStrength.BenignSupporting,
+                    summary="BP3 is not applicable for PALB2.",
+                ),
+            )
+        return super().predict_pm4bp3(seqvar, var_data)
+
     def predict_bp7(self, seqvar: SeqVar, var_data: AutoACMGData) -> AutoACMGCriteria:
         """Override donor and acceptor positions for ATM and PALB2."""
         if var_data.hgnc_id == "HGNC:26144":