feat: download alphamissense scores

config/config.yaml

@@ -206,10 +206,7 @@ annotations:
         # For integration of SpliceAI preprocessed snv and indel scores (including index files) are required. 
         # Scores are available on https://basespace.illumina.com/s/otSPW8hnhaZR
         #- SpliceAI,snv=<path/to/spliceai_scores.raw.snv.hg38.vcf.gz,indel=<path/to/spliceai_scores.raw.indel.hg38.vcf.gz>
-        # For annotation of AlphaMissense scores a tsv-file containing processed scores is required.
-        # Scores are available on https://zenodo.org/records/10813168
-        # A tabix index is required and can be created by calling `tabix -s 1 -b 2 -e 2 -f -S 1 AlphaMissense_hg38.tsv.gz`
-        #- AlphaMissense,file=<path/to/AlphaMissense_hg38.tsv.gz>
+        #- AlphaMissense
 
 # printing of variants in a table format (might be deprecated soon)
 tables:

workflow/resources/datavzrd/variant-calls-template.datavzrd.yaml

@@ -388,6 +388,7 @@ views:
               ensembl:
                 url: https://www.ensembl.org/Homo_sapiens/Transcript/Summary?t={feature}
           hgvsc:
+            optional: true
             custom: ?read_file(input.linkouts)
           consequence:
             plot:

workflow/rules/annotation.smk

@@ -29,6 +29,8 @@ rule annotate_variants:
         plugins="resources/vep/plugins",
         revel=lambda wc: get_plugin_aux("REVEL"),
         revel_tbi=lambda wc: get_plugin_aux("REVEL", True),
+        alphamissense=lambda wc: get_plugin_aux("AlphaMissense"),
+        alphamissense_tbi=lambda wc: get_plugin_aux("AlphaMissense", True),
         fasta=genome,
         fai=genome_fai,
     output:

workflow/rules/common.smk

@@ -5,6 +5,10 @@ import yaml
 import pandas as pd
 from snakemake.utils import validate
 
+
+ruleorder: tabix_plugin_scores > tabix_known_variants
+
+
 validate(config, schema="../schemas/config.schema.yaml")
 
 samples = (
@@ -516,7 +520,7 @@ def get_markduplicates_extra(wc):
 def get_group_bams(wildcards, bai=False):
     ext = "bai" if bai else "bam"
     if is_activated("primers/trimming") and not group_is_paired_end(wildcards.group):
-        WorkflowError("Primer trimming is only available for paired end data.")
+        raise WorkflowError("Primer trimming is only available for paired end data.")
     return expand(
         "results/recal/{sample}.{ext}",
         sample=get_group_samples(wildcards.group),
@@ -725,9 +729,12 @@ def get_vep_threads():
 
 def get_plugin_aux(plugin, index=False):
     if plugin in config["annotations"]["vep"]["final_calls"]["plugins"]:
+        suffix = ".tbi" if index else ""
         if plugin == "REVEL":
             suffix = ".tbi" if index else ""
             return "resources/revel_scores.tsv.gz{suffix}".format(suffix=suffix)
+        if plugin == "AlphaMissense":
+            return "resources/alphamissense_scores.tsv.gz{suffix}".format(suffix=suffix)
     return []
 
 
@@ -982,10 +989,27 @@ def get_tabix_params(wildcards):
     raise ValueError("Invalid format for tabix: {}".format(wildcards.format))
 
 
-def get_tabix_revel_params():
-    # Indexing of REVEL-score file where the column depends on the reference
-    column = 2 if config["ref"]["build"] == "GRCh37" else 3
-    return f"-f -s 1 -b {column} -e {column}"
+def get_tabix_plugin_params(plugin):
+    if plugin == "revel":
+        # Indexing of REVEL-score file where the column depends on the reference
+        column = 2 if config["ref"]["build"] == "GRCh37" else 3
+        return f"-f -s 1 -b {column} -e {column}"
+    elif plugin == "alphamissense":
+        return "-f -s 1 -b 2 -e 2 -f -S 1"
+    else:
+        raise WorkflowError("Unsupported plugin for obtaining tabix parameteres")
+
+
+def get_alphamissense_url(wc):
+    if config["ref"]["build"] == "GRCh37":
+        build = "hg19"
+    elif config["ref"]["build"] == "GRCh38":
+        build = "hg38"
+    else:
+        raise WorkflowError(
+            "Invalid reference for AlphaMissense annotation. Only GRCh37 and GRCh38 supported."
+        )
+    return f"https://zenodo.org/records/10813168/files/AlphaMissense_{build}.tsv.gz"
 
 
 def get_untrimmed_fastqs(wc):

workflow/rules/plugins.smk

@@ -35,12 +35,27 @@ rule process_revel_scores:
         """
 
 
-use rule tabix_known_variants as tabix_revel_scores with:
+rule download_alphamissense_scores:
+    output:
+        "resources/alphamissense_scores.tsv.gz",
+    params:
+        url=get_alphamissense_url,
+    log:
+        "logs/vep_plugins/download_alphamissense.log",
+    conda:
+        "../envs/curl.yaml"
+    shell:
+        """
+        curl {params.url} -o {output} &> {log}
+        """
+
+
+use rule tabix_known_variants as tabix_plugin_scores with:
     input:
-        "resources/revel_scores.tsv.gz",
+        "resources/{plugin}_scores.tsv.gz",
     output:
-        "resources/revel_scores.tsv.gz.tbi",
+        "resources/{plugin}_scores.tsv.gz.tbi",
     params:
-        get_tabix_revel_params(),
+        lambda wc: get_tabix_plugin_params(wc.plugin),
     log:
-        "logs/tabix/revel.log",
+        "logs/tabix/{plugin}.log",