Init pp3 and bp4 implementation (#115)

* init pp3 and bp4

* add empty test

docs/acmg_seqvars_implementation.rst

@@ -72,3 +72,14 @@ BP7
     :undoc-members:
     :show-inheritance:
     :private-members:
+
+-----------
+PP3 and BP4
+-----------
+
+.. automodule:: src.criteria.auto_pp3_bp4
+    :members:
+    :inherited-members:
+    :undoc-members:
+    :show-inheritance:
+    :private-members:

src/auto_acmg.py

@@ -16,17 +16,17 @@
 
 #: Criteria, which have no automatic prediction yet.
 NOT_IMPLEMENTED_CRITERIA = [
-    "PS2",   # De novo (both parents not tested)
-    "PS3",   # Well-established in vitro or in vivo functional studies
-    "PS4",   # Increased prevalence in affected individuals vs. controls
-    "PM3",   # Recessive disorder (zygosity unknown)
-    "PM6",   # Assumed de novo (both parents not tested)
-    "PP1",   # Cosegregation with disease in multiple affected family members
-    "PP4",   # Patient's phenotype or family history specificity
-    "BS3",   # Well-established in vitro or in vivo functional studies
-    "BS4",   # Lack of segregation in affected family members
-    "BP2",   # Observed in healthy individuals (zygosity unknown)
-    "BP5",   # Case with an alternate molecular basis for disease
+    "PS2",  # De novo (both parents not tested)
+    "PS3",  # Well-established in vitro or in vivo functional studies
+    "PS4",  # Increased prevalence in affected individuals vs. controls
+    "PM3",  # Recessive disorder (zygosity unknown)
+    "PM6",  # Assumed de novo (both parents not tested)
+    "PP1",  # Cosegregation with disease in multiple affected family members
+    "PP4",  # Patient's phenotype or family history specificity
+    "BS3",  # Well-established in vitro or in vivo functional studies
+    "BS4",  # Lack of segregation in affected family members
+    "BP2",  # Observed in healthy individuals (zygosity unknown)
+    "BP5",  # Case with an alternate molecular basis for disease
 ]
 
 

src/criteria/auto_acmg.py

@@ -9,6 +9,7 @@
 from src.criteria.auto_pm1 import AutoPM1
 from src.criteria.auto_pm4_bp3 import AutoPM4BP3
 from src.criteria.auto_pp2_bp1 import AutoPP2BP1
+from src.criteria.auto_pp3_bp4 import AutoPP3BP4
 from src.criteria.auto_ps1_pm5 import AutoPS1PM5
 from src.defs.annonars_variant import AnnonarsVariantResponse
 from src.defs.auto_acmg import ACMGCriteria
@@ -139,4 +140,19 @@ def predict(self) -> Optional[ACMGCriteria]:
         except AutoAcmgBaseException as e:
             logger.error("Failed to predict BP7 criteria. Error: {}", e)
 
+        # PP3 and BP4
+        try:
+            logger.info("Predicting PP3 and BP4 criteria.")
+            pp3_bp4 = AutoPP3BP4(
+                self.seqvar, self.genome_release, variant_info.result, config=self.config
+            )
+            pp3_bp4_prediction = pp3_bp4.predict()
+            if not pp3_bp4_prediction:
+                logger.error("Failed to predict PP3 and BP4 criteria.")
+            else:
+                self.prediction.PP3 = pp3_bp4_prediction.PP3
+                self.prediction.BP4 = pp3_bp4_prediction.BP4
+        except AutoAcmgBaseException as e:
+            logger.error("Failed to predict PP3 and BP4 criteria. Error: {}", e)
+
         return self.prediction

src/criteria/auto_pp3_bp4.py

@@ -0,0 +1,110 @@
+"""Implementation of the PP3 and BP4 criteria."""
+
+from typing import List, Optional
+
+from loguru import logger
+
+from src.api.annonars import AnnonarsClient
+from src.core.config import Config
+from src.defs.annonars_variant import VariantResult
+from src.defs.auto_acmg import PP3BP4
+from src.defs.exceptions import AutoAcmgBaseException, MissingDataError
+from src.defs.genome_builds import GenomeRelease
+from src.defs.seqvar import SeqVar
+
+
+class AutoPP3BP4:
+    """Class for automatic PP3 and BP4 prediction."""
+
+    def __init__(
+        self,
+        seqvar: SeqVar,
+        genome_release: GenomeRelease,
+        variant_info: VariantResult,
+        *,
+        config: Config,
+    ):
+        """Initialize the class."""
+        self.seqvar = seqvar
+        self.genome_release = genome_release
+        self.variant_info = variant_info
+        self.config = config
+        self.annonars_client = AnnonarsClient(api_base_url=config.api_base_url_annonars)
+        self.prediction: PP3BP4 | None = None
+
+    @staticmethod
+    def _get_best_pathogenic_score(variant_info: VariantResult) -> Optional[float]:
+        """Get the best pathogenic score available."""
+        # Avoid linting errors
+        assert variant_info.cadd is not None and variant_info.dbnsfp is not None
+        scores = [
+            # variant_info.mutpred2,
+            variant_info.cadd.PolyPhenVal,
+            # variant_info.revel,
+            # variant_info.bayesdel,
+            # variant_info.vest4,
+            # variant_info.phylop,
+        ]
+        scores = [score for score in scores if score is not None]
+        return max(scores) if scores else None  # type: ignore
+
+    @staticmethod
+    def _get_best_benign_score(variant_info: VariantResult) -> Optional[float]:
+        """Get the best benign score available."""
+        # Avoid linting errors
+        assert variant_info.cadd is not None and variant_info.dbnsfp is not None
+        scores = [
+            # variant_info.dbnsfp.REVEL_rankscore,
+            # variant_info.dbnsfp.MutPred_rankscore,
+            variant_info.cadd.PolyPhenVal,
+            # variant_info.bayesdel,
+            # variant_info.vest4,
+            # variant_info.phylop,
+        ]
+        scores = [score for score in scores if score is not None]
+        return min(scores) if scores else None  # type: ignore
+
+    def _predict_spliceai(self, variant_info: VariantResult) -> Optional[float]:
+        """Predict splice site alterations using SpliceAI."""
+        # TODO: Implement this method.
+        return None
+
+    def predict(self) -> Optional[PP3BP4]:
+        """Predict PP3 and BP4 criteria."""
+        self.prediction = PP3BP4()
+
+        try:
+            if self.seqvar.chrom == "MT":
+                self.prediction.PP3 = False
+                self.prediction.BP4 = False
+                return self.prediction
+
+            if not self.variant_info or not self.variant_info.cadd or not self.variant_info.dbnsfp:
+                logger.error("Missing CADD or DBNSFP data.")
+                raise MissingDataError("Missing CADD or DBNSFP data.")
+
+            best_pathogenic_score = self._get_best_pathogenic_score(self.variant_info)
+            best_benign_score = self._get_best_benign_score(self.variant_info)
+            spliceai_score = self._predict_spliceai(self.variant_info)
+
+            # Evaluate PP3
+            if best_pathogenic_score and best_pathogenic_score > 0.8:
+                self.prediction.PP3 = True
+            elif spliceai_score and spliceai_score > 0.8:
+                self.prediction.PP3 = True
+            else:
+                self.prediction.PP3 = False
+
+            # Evaluate BP4
+            if best_benign_score and best_benign_score < 0.2:
+                self.prediction.BP4 = True
+            elif spliceai_score and spliceai_score < 0.2:
+                self.prediction.BP4 = True
+            else:
+                self.prediction.BP4 = False
+
+        except AutoAcmgBaseException as e:
+            logger.error("Failed to predict PP3 and BP4 criteria. Error: {}", e)
+            self.prediction = None
+
+        return self.prediction

src/defs/annonars_variant.py

@@ -19,6 +19,7 @@ class VariantQuery(BaseModel):
 
 class Cadd(BaseModel):
     ConsDetail: Optional[str] = None
+    PolyPhenVal: Optional[float] = None
 
 
 class GnomadExomes(BaseModel):

src/defs/auto_acmg.py

@@ -122,3 +122,10 @@ class BP7(BaseModel):
     """BP7 criteria prediction."""
 
     BP7: bool = False
+
+
+class PP3BP4(BaseModel):
+    """PP3 and BP4 criteria prediction."""
+
+    PP3: bool = False
+    BP4: bool = False

tests/criteria/test_auto_ba1_bs1_bs2_pm2.py

@@ -0,0 +1 @@
+pass

tests/criteria/test_auto_bp7.py

@@ -0,0 +1 @@
+pass

tests/criteria/test_auto_pp2_bp1.py

@@ -0,0 +1 @@
+pass

tests/criteria/test_auto_pp3_bp4.py

@@ -0,0 +1 @@
+pass