New options for transcripts prioritization #256

howard/objects/variants.py

@@ -1091,7 +1091,6 @@ def get_header(self, type: str = "vcf"):
  elif type == "list":
  return vcf_required
 
-
  def get_header_infos_list(self) -> list:
  """
  This function retrieves a list of information fields from the header.
@@ -1106,7 +1105,6 @@ def get_header_infos_list(self) -> list:
 
  return infos_list
 
-
  def get_header_length(self, file: str = None) -> int:
  """
  The function `get_header_length` returns the length of the header list, excluding the #CHROM
@@ -9750,29 +9748,10 @@ def transcripts_prioritization(
  table=transcripts_table,
  pz_param=param.get("transcripts", {}).get("prioritization", {}),
  )
-
- # # DEVEL
- # query_devel = f"""
- # SELECT * FROM transcripts LIMIT 1
- # """
- # res_devel = self.get_query_to_df(query=query_devel)
- # log.debug(f""" res_devel transcripts ={res_devel.to_string()} """)
- # query_devel = f"""
- # SELECT INFO FROM variants LIMIT 1
- # """
- # res_devel = self.get_query_to_df(query=query_devel)
- # log.debug(f""" res_devel variants ={res_devel.to_string()} """)
-
  if not prioritization_result:
  log.warning("Transcripts prioritization not processed")
  return False
 
- # # Explode PZ fields
- # self.explode_infos(
- # table=transcripts_table,
- # fields=list(pz_param_pzfields.keys()) + pz_mandatory_fields,
- # )
-
  # PZ fields sql query
  query_update_select_list = []
  query_update_concat_list = []
@@ -9795,70 +9774,131 @@ def transcripts_prioritization(
 
  # Order by
  pz_orders = (
- param.get("transcripts", {}).get("prioritization", {}).get("pzorder", {})
+ param.get("transcripts", {})
+ .get("prioritization", {})
+ .get("prioritization_transcripts_order", {})
  )
  if not pz_orders:
  pz_orders = {
  pz_param.get("pzprefix", "PTZ") + "Flag": "ASC",
  pz_param.get("pzprefix", "PTZ") + "Score": "DESC",
- "transcript": "ASC",
  }
  for pz_order in pz_orders:
  query_update_order_list.append(
  f""" {pz_order} {pz_orders.get(pz_order, "DESC")} """
  )
 
  # Fields to explode
- fields_to_explode = list(pz_param_pzfields.keys()) + pz_mandatory_fields + list(pz_orders.keys())
- fields_to_explode.remove("transcript")
+ fields_to_explode = (
+ list(pz_param_pzfields.keys())
+ + pz_mandatory_fields
+ + list(pz_orders.keys())
+ )
+ # Remove transcript column as a specific transcript column
+ if "transcript" in fields_to_explode:
+ fields_to_explode.remove("transcript")
 
  # Check fields to explode
  for field_to_explode in fields_to_explode:
  if field_to_explode not in self.get_header_infos_list():
  msg_err = f"INFO/{field_to_explode} NOT IN header"
  log.error(msg_err)
  raise ValueError(msg_err)
- 
- # eExplode fields to explode
- exploded_fields = self.explode_infos(
+
+ # Explode fields to explode
+ self.explode_infos(
  table=transcripts_table,
  fields=fields_to_explode,
  )
-
- # # DEVEL
- # query_devel = f"""
- # DESCRIBE 
-
- # {transcripts_table}
- # """
- # log.debug(f""" query_devel={query_devel} """)
- # df_devel = self.get_query_to_df(query=query_devel)
- # log.debug(df_devel)
- # query_devel = f"""
- # SELECT
- # "#CHROM", POS, REF, ALT, transcript, {" ".join(query_update_select_list)}
- # ROW_NUMBER() OVER (
- # PARTITION BY "#CHROM", POS, REF, ALT
- # ORDER BY {" , ".join(query_update_order_list)}
- # ) AS rn
- # FROM
- # {transcripts_table}
- # """
- # log.debug(f""" query_devel={query_devel} """)
- # df_devel = self.get_query_to_df(query=query_devel)
- # log.debug(df_devel)
 
- # Export Transcripts prioritization infos to variants table
- query_update = f"""
- WITH RankedTranscripts AS (
+ # Transcript preference file
+ transcripts_preference_file = (
+ param.get("transcripts", {})
+ .get("prioritization", {})
+ .get("prioritization_transcripts", {})
+ )
+ transcripts_preference_file = full_path(transcripts_preference_file)
+
+ # Transcript preference forced
+ transcript_preference_force = (
+ param.get("transcripts", {})
+ .get("prioritization", {})
+ .get("prioritization_transcripts_force", False)
+ )
+ # Transcript version forced
+ transcript_version_force = (
+ param.get("transcripts", {})
+ .get("prioritization", {})
+ .get("prioritization_transcripts_version_force", False)
+ )
+
+ # Transcripts Ranking
+ if transcripts_preference_file:
+
+ # Transcripts file to dataframe
+ if os.path.exists(transcripts_preference_file):
+ transcripts_preference_dataframe = transcripts_file_to_df(
+ transcripts_preference_file
+ )
+ else:
+ log.error(
+ f"Transcript file '{transcripts_preference_file}' does NOT exist"
+ )
+ raise ValueError(
+ f"Transcript file '{transcripts_preference_file}' does NOT exist"
+ )
+
+ # Order by depending to transcript preference forcing
+ if transcript_preference_force:
+ order_by = f""" transcripts_preference.transcripts_preference_order ASC, {" , ".join(query_update_order_list)} """
+ else:
+ order_by = f""" {" , ".join(query_update_order_list)}, transcripts_preference.transcripts_preference_order ASC """
+
+ # Transcript columns joined depend on version consideration
+ if transcript_version_force:
+ transcripts_version_join = f""" {transcripts_table}.transcript = transcripts_preference.transcripts_preference """
+ else:
+ transcripts_version_join = f""" split_part({transcripts_table}.transcript, '.', 1) = split_part(transcripts_preference.transcripts_preference, '.', 1) """
+
+ # Query ranking for update
+ query_update_ranking = f"""
+ SELECT
+ "#CHROM", POS, REF, ALT, {transcripts_table}.transcript, {" ".join(query_update_select_list)}
+ ROW_NUMBER() OVER (
+ PARTITION BY "#CHROM", POS, REF, ALT
+ ORDER BY {order_by}
+ ) AS rn
+ FROM {transcripts_table}
+ LEFT JOIN 
+ (
+ SELECT transcript AS 'transcripts_preference', row_number() OVER () AS transcripts_preference_order
+ FROM transcripts_preference_dataframe
+ ) AS transcripts_preference
+ ON {transcripts_version_join}
+ """
+
+ else:
+
+ # Query ranking for update
+ query_update_ranking = f"""
  SELECT
  "#CHROM", POS, REF, ALT, transcript, {" ".join(query_update_select_list)}
  ROW_NUMBER() OVER (
  PARTITION BY "#CHROM", POS, REF, ALT
  ORDER BY {" , ".join(query_update_order_list)}
  ) AS rn
- FROM
- {transcripts_table}
+ FROM {transcripts_table}
+ """
+
+ # DEBUG
+ # log.debug(f""" query_update_ranking={query_update_ranking} """)
+ # df_devel = self.get_query_to_df(query=query_update_ranking)
+ # log.debug(df_devel)
+
+ # Export Transcripts prioritization infos to variants table
+ query_update = f"""
+ WITH RankedTranscripts AS (
+ {query_update_ranking}
  )
  UPDATE {table_variants}
  SET
@@ -9876,9 +9916,9 @@ def transcripts_prioritization(
  AND variants."#CHROM" = RankedTranscripts."#CHROM"
  AND variants."POS" = RankedTranscripts."POS"
  AND variants."REF" = RankedTranscripts."REF"
- AND variants."ALT" = RankedTranscripts."ALT"
- 
+ AND variants."ALT" = RankedTranscripts."ALT" 
  """
+
  # log.debug(f"query_update={query_update}")
  self.execute_query(query=query_update)
 

tests/data/prioritization_transcripts_profiles.json

@@ -0,0 +1,54 @@
+{
+ "transcripts": {
+ "LIST_S2_score": [
+ {
+ "type": "gt",
+ "value": "0.75",
+ "score": 10,
+ "flag": "PASS",
+ "comment": ["Very Good LIST Score"]
+ },
+ {
+ "type": "gt",
+ "value": "0.50",
+ "score": 10,
+ "flag": "PASS",
+ "comment": ["Good LIST Score"]
+ }
+ ],
+ "CLNSIG": [
+ {
+ "type": "eq",
+ "value": "pathogenic",
+ "score": 100,
+ "flag": "PASS",
+ "comment": ["Pathogenic"]
+ }
+ ],
+ "AnnotationImpact": [
+ {
+ "type": "eq",
+ "value": "MODIFIER",
+ "score": 100,
+ "flag": "PASS",
+ "comment": ["MODIFIER"]
+ }
+ ],
+ "transcript": [
+ {
+ "type": "eq",
+ "value": "NM_005228.5",
+ "score": 100,
+ "flag": "PASS",
+ "comment": ["NM_005228.5"]
+ },
+ {
+ "type": "eq",
+ "value": "NM_001346941.2",
+ "score": 100,
+ "flag": "PASS",
+ "comment": ["NM_001346941.2"]
+ }
+ ]
+ }
+}

tests/data/transcripts.for_prioritization.tsv

@@ -0,0 +1,6 @@
+NR_024540 WASH7P
+NR_036266 MIR1302-9
+NM_001346900 EGFR
+NM_001346897 EGFR
+NR_047551 EGFR
+NM_001346941.2 EGFR