prefilter.py

prefilter.py

@@ -0,0 +1,70 @@
+"""prefilter.py — PeVe v1.1"""
+from __future__ import annotations
+from dataclasses import dataclass, field
+from typing import Optional
+from config import VEP_CONSEQUENCE_MAP, L3_SUBSTITUTION_INVALID
+
+@dataclass
+class VariantClass:
+    raw_consequence: str
+    all_consequences: list
+    variant_class: str
+    l3_substitution_valid: bool
+    rna_priority: bool
+    protein_priority: bool
+    protein_deprioritised: bool
+    transcript_conflict: bool
+    out_of_scope: bool
+    flags: list = field(default_factory=list)
+
+def classify_variant(ref, alt, vep_consequence, all_vep_consequences=None):
+    if all_vep_consequences is None:
+        all_vep_consequences = [vep_consequence]
+    cons     = vep_consequence.lower().strip()
+    all_cons = [c.lower().strip() for c in all_vep_consequences]
+
+    # MNV detection
+    if len(ref) > 1 and len(alt) > 1 and len(ref) == len(alt):
+        return VariantClass(cons, all_cons, "mnv", False, False, False, False, False, True,
+                            ["MNV: single-variant assessment may be incomplete"])
+
+    variant_class = VEP_CONSEQUENCE_MAP.get(cons, "unknown")
+    if variant_class == "unknown":
+        variant_class = _infer(ref, alt)
+
+    mapped = {VEP_CONSEQUENCE_MAP.get(c, "unknown") for c in all_cons}
+    tx_conflict = len(mapped) > 1
+
+    l3_valid             = variant_class not in L3_SUBSTITUTION_INVALID
+    rna_priority         = variant_class == "canonical_splice"
+    protein_priority     = variant_class == "substitution_missense"
+    protein_deprio       = variant_class == "substitution_synonymous"
+    out_of_scope         = variant_class in {"utr_regulatory", "mnv", "unknown"}
+
+    flags = []
+    if variant_class == "utr_regulatory":
+        flags.append("UTR/regulatory: no mechanism pathway in PeVe v1.1.")
+    if variant_class in {"frameshift", "stop_gained", "start_lost"}:
+        flags.append(f"{variant_class}: Layer 3 substitution metrics NOT APPLICABLE.")
+    if variant_class == "in_frame_indel":
+        flags.append("In-frame indel: substitution biochemistry NOT APPLICABLE.")
+    if variant_class == "deep_intronic":
+        flags.append("Deep intronic: RNA interpretation down-prioritised.")
+    if variant_class == "substitution_synonymous":
+        flags.append("Synonymous: context signal alone cannot classify pathogenic.")
+    if tx_conflict:
+        flags.append("Transcript conflict: consequence differs across transcripts.")
+    if variant_class == "unknown":
+        flags.append("Variant class unknown — outputs are exploratory only.")
+
+    return VariantClass(cons, all_cons, variant_class, l3_valid,
+                        rna_priority, protein_priority, protein_deprio,
+                        tx_conflict, out_of_scope, flags)
+
+def _infer(ref, alt):
+    if len(ref) == 1 and len(alt) == 1:
+        return "substitution_missense"
+    diff = len(alt) - len(ref)
+    if diff % 3 == 0:
+        return "in_frame_indel"
+    return "frameshift"