Merge pull request #114 from bcgsc/feature/GERO-296_tmbur_high_matches

Feature/gero 296 tmbur high matches

ipr/annotate.py

@@ -11,10 +11,11 @@
 from graphkb.match import INPUT_COPY_CATEGORIES
 from graphkb.types import Variant
 from graphkb.util import FeatureNotFoundError, convert_to_rid_list
+from pandas import isnull
 from progressbar import progressbar
 from typing import Any, Dict, List, Sequence, Set, cast
 
-from .constants import FAILED_REVIEW_STATUS
+from .constants import FAILED_REVIEW_STATUS, TMB_HIGH_CATEGORY
 from .ipr import convert_statements_to_alterations
 from .types import (
     GkbStatement,
@@ -328,7 +329,8 @@ def annotate_positional_variants(
 
         for var_key in VARIANT_KEYS:
             variant = row.get(var_key)
-            if not variant:
+            matches = []
+            if not variant or isnull(variant):
                 continue
             try:
                 try:
@@ -349,12 +351,14 @@ def annotate_positional_variants(
                         matches = gkb_match.match_positional_variant(graphkb_conn, variant)
 
                 # GERO-299 - check for conflicting nonsense and missense categories
+
                 missense = [
                     m for m in matches if 'missense' in m.get('type', m).get('displayName', '')
                 ]
                 nonsense = [
                     m for m in matches if 'nonsense' in m.get('type', m).get('displayName', '')
                 ]
+
                 missense_cat = [m for m in missense if m.get('@class', '') == 'CategoryVariant']
                 nonsense_cat = [m for m in nonsense if m.get('@class', '') == 'CategoryVariant']
                 if missense_cat and nonsense_cat:
@@ -399,6 +403,9 @@ def annotate_positional_variants(
                         matches = [
                             m for m in matches if m not in missense_cat and m not in nonsense_cat
                         ]
+                elif nonsense_cat and ':c.' in variant:
+                    logger.error(f"GERO-304 - dropping nonsense variants from hgvsCds {variant}")
+                    matches = [m for m in matches if m not in nonsense_cat]
 
                 for ipr_row in get_ipr_statements_from_variants(
                     graphkb_conn, matches, disease_name
@@ -444,8 +451,8 @@ def annotate_positional_variants(
 
 def annotate_msi(
     graphkb_conn: GraphKBConnection,
-    msi_category: str,
-    disease_name: str,
+    disease_name: str = 'cancer',
+    msi_category: str = 'microsatellite instability',
 ) -> List[KbMatch]:
     """Annotate microsatellite instablity from GraphKB in the IPR alterations format.
 
@@ -476,3 +483,43 @@ def annotate_msi(
             ipr_row['variantType'] = 'msi'
             gkb_matches.append(ipr_row)
     return gkb_matches
+
+
+def annotate_tmb(
+    graphkb_conn: GraphKBConnection,
+    disease_name: str = 'cancer',
+    category: str = TMB_HIGH_CATEGORY,
+) -> List[KbMatch]:
+    """Annotate Tumour Mutation Burden (tmb) categories from GraphKB in the IPR alterations format.
+
+    Match to GraphKb Category variants with similar names
+    Args:
+        graphkb_conn: the graphkb api connection object
+        disease_name: oncotree disease name for graphkb matching.
+        category: such as 'high mutation burden'
+
+    Returns:
+        list of kbMatches records for IPR
+    """
+    gkb_matches = []
+    categories = graphkb_conn.query(
+        {
+            'target': {
+                'target': 'CategoryVariant',
+                'filters': {
+                    'reference1': {
+                        'target': 'Signature',
+                        'filters': {'OR': [{'name': category}, {'displayName': category}]},
+                    }
+                },
+            },
+            'queryType': 'similarTo',
+            'returnProperties': ['@rid', 'displayName'],
+        },
+    )
+    if categories:
+        for ipr_row in get_ipr_statements_from_variants(graphkb_conn, categories, disease_name):
+            ipr_row['variant'] = category
+            ipr_row['variantType'] = 'tmb'
+            gkb_matches.append(ipr_row)
+    return gkb_matches

ipr/constants.py

@@ -4,3 +4,6 @@
 
 # all possible values for review status are: ['pending', 'not required', 'passed', 'failed', 'initial']
 FAILED_REVIEW_STATUS = 'failed'
+
+TMB_HIGH = 10.0  # genomic mutations per mb
+TMB_HIGH_CATEGORY = 'high mutation burden'

ipr/content.spec.json

@@ -1197,6 +1197,22 @@
                             "null"
                         ]
                     },
+                    "rnaAltCount": {
+                        "description": "the number of alternate reads in the rna supporting the mutation",
+                        "example": 1,
+                        "type": [
+                            "integer",
+                            "null"
+                        ]
+                    },
+                    "rnaDepth": {
+                        "description": "the total number of reads at this position in the rna",
+                        "example": 2,
+                        "type": [
+                            "integer",
+                            "null"
+                        ]
+                    },
                     "svg": {
                         "description": "svg image file content for this SV",
                         "type": [

ipr/inputs.py

@@ -139,7 +139,8 @@
     'highQuality',
     'comments',
     'library',
-    # GERO-307 - tumourAltCount and tumourDepth are available but not rnaAltCount and rnaDepth
+    'rnaAltCount',
+    'rnaDepth',
     'tumourAltCount',
     'tumourDepth',
     'germline',

ipr/main.py

@@ -12,10 +12,11 @@
     annotate_expression_variants,
     annotate_msi,
     annotate_positional_variants,
+    annotate_tmb,
     get_gene_information,
 )
 from .connection import IprConnection
-from .constants import DEFAULT_URL
+from .constants import DEFAULT_URL, TMB_HIGH, TMB_HIGH_CATEGORY
 from .inputs import (
     check_comparators,
     check_variant_links,
@@ -198,10 +199,38 @@ def create_report(
     gkb_matches: List[KbMatch] = []
 
     # Signature category variants
+    tmb_variant: IprVariant = {}
+    tmb_matches = []
     if 'tmburMutationBurden' in content.keys():
-        logger.warning(
-            'GERO-296 - not yet implemented - high tumour mutation burden category matching.'
-        )
+        tmb_val = 0.0
+        tmb = {}
+        try:
+            tmb = content.get('tmburMutationBurden', {})
+            tmb_val = tmb['genomeIndelTmb'] + tmb['genomeSnvTmb']
+        except Exception as err:
+            logger.error(f"tmburMutationBurden parsing failure: {err}")
+
+        if tmb_val >= TMB_HIGH:
+            logger.warning(
+                f'GERO-296 - tmburMutationBurden high -checking graphkb matches for {TMB_HIGH_CATEGORY}'
+            )
+            if not tmb.get('key'):
+                tmb['key'] = TMB_HIGH_CATEGORY
+            if not tmb.get('kbCategory'):
+                tmb['kbCategory'] = TMB_HIGH_CATEGORY
+
+            # GERO-296 - try matching to graphkb
+            tmb_matches = annotate_tmb(graphkb_conn, kb_disease_match, TMB_HIGH_CATEGORY)
+            if tmb_matches:
+                tmb_variant['kbCategory'] = TMB_HIGH_CATEGORY  # type: ignore
+                tmb_variant['variant'] = TMB_HIGH_CATEGORY
+                tmb_variant['key'] = tmb['key']
+                tmb_variant['variantType'] = 'tmb'
+                logger.info(
+                    f"GERO-296 '{TMB_HIGH_CATEGORY}' matches {len(tmb_matches)} statements."
+                )
+                gkb_matches.extend(tmb_matches)
+                logger.debug(f"\tgkb_matches: {len(gkb_matches)}")
 
     msi = content.get('msi', [])
     msi_matches = []
@@ -216,7 +245,7 @@ def create_report(
             msi_cat = msi.get('kbCategory')
             msi_variant = msi.copy()
         logger.info(f'Matching GKB msi {msi_cat}')
-        msi_matches = annotate_msi(graphkb_conn, msi_cat, kb_disease_match)
+        msi_matches = annotate_msi(graphkb_conn, kb_disease_match, msi_cat)
         if msi_matches:
             msi_variant['kbCategory'] = msi_cat  # type: ignore
             msi_variant['variant'] = msi_cat
@@ -262,6 +291,8 @@ def create_report(
     all_variants = expression_variants + copy_variants + structural_variants + small_mutations  # type: ignore
     if msi_matches:
         all_variants.append(msi_variant)  # type: ignore
+    if tmb_matches:
+        all_variants.append(tmb_variant)  # type: ignore
 
     if match_germline:  # verify germline kb statements matched germline observed variants
         gkb_matches = germline_kb_matches(gkb_matches, all_variants)

tests/test_annotate.py

@@ -102,7 +102,6 @@ def test_annotate_structural_variants_tp53(graphkb_conn):
     disease = 'cancer'
     ref_key = 'prot_only'
     pref = annotate_positional_variants(graphkb_conn, [TP53_MUT_DICT[ref_key]], disease)
-    known_issues = set(['TP53:p.M237X'])  # SDEV-3122 -
     # GERO-299 - nonsense - stop codon - should not match.  This is missense not nonsense (#164:933).
     nonsense = [a for a in pref if a['kbVariant'] == 'TP53 nonsense']
     assert not nonsense
@@ -117,15 +116,11 @@ def test_annotate_structural_variants_tp53(graphkb_conn):
         diff = pref_vars.symmetric_difference(alt_vars)
         missing = pref_vars.difference(alt_vars)
 
-        known_issues = set([])
-        if 'hgvsCds' in alt_rep:
-            known_issues.add('TP53 nonsense')  # GERO-299
-        if 'p.M237' not in alt_rep:
-            known_issues.add('TP53:p.M237X')  # SDEV-3122 - not matching imprecise mutations
+        known_issues = set()
         if key == 'genome_only':
+            # genome_only matched to more precise type 'TP53 deleterious mutation' but not 'TP53 mutation'
             known_issues.add('TP53 mutation')
 
-        # strangely genome_only matched to more precise type 'TP53 deleterious mutation' but not 'TP53 mutation'
         missing = pref_vars.difference(alt_vars).difference(known_issues)
         print(alt_vars)
         assert not missing, f"{key} missing{missing}: {diff}"