wip

snappy_pipeline/apps/snappy_snake.py

@@ -411,7 +411,7 @@ def main(argv=None):
         dest="conda_create_envs_only",
         action="store_true",
         default=False,
-        help="Prepare all conda environments"
+        help="Prepare all conda environments",
     )
 
     args = parser.parse_args(argv)

snappy_pipeline/workflows/variant_calling/Snakefile

@@ -73,23 +73,39 @@ rule variant_calling_bcftools_run:
 # Run GATK HaplotypeCaller (direct, pedigree/multi-sample) --------------------
 
 
-rule variant_calling_gatk_hc_run:
+rule variant_calling_gatk3_hc_run:
     input:
-        unpack(wf.get_input_files("gatk_hc", "run")),
+        unpack(wf.get_input_files("gatk3_hc", "run")),
     output:
-        **wf.get_output_files("gatk_hc", "run"),
-    threads: wf.get_resource("gatk_hc", "run", "threads")
+        **wf.get_output_files("gatk3_hc", "run"),
+    threads: wf.get_resource("gatk3_hc", "run", "threads")
     resources:
-        time=wf.get_resource("gatk_hc", "run", "time"),
-        memory=wf.get_resource("gatk_hc", "run", "memory"),
-        partition=wf.get_resource("gatk_hc", "run", "partition"),
-    params:
-        step_key="variant_calling",
-        caller_key="gatk_hc",
+        time=wf.get_resource("gatk3_hc", "run", "time"),
+        memory=wf.get_resource("gatk3_hc", "run", "memory"),
+        partition=wf.get_resource("gatk3_hc", "run", "partition"),
+    log:
+        **wf.get_log_file("gatk3_hc", "run"),
+    wrapper:
+        wf.wrapper_path("gatk3_hc")
+
+
+# Run GATK UnifiedGenotyper ---------------------------------------------------
+
+
+rule variant_calling_gatk3_ug_run:
+    input:
+        unpack(wf.get_input_files("gatk3_ug", "run")),
+    output:
+        **wf.get_output_files("gatk3_ug", "run"),
+    threads: wf.get_resource("gatk3_ug", "run", "threads")
+    resources:
+        time=wf.get_resource("gatk3_ug", "run", "time"),
+        memory=wf.get_resource("gatk3_ug", "run", "memory"),
+        partition=wf.get_resource("gatk3_ug", "run", "partition"),
     log:
-        **wf.get_log_file("gatk_hc", "run"),
+        **wf.get_log_file("gatk3_ug", "run"),
     wrapper:
-        wf.wrapper_path("gatk_hc_par")
+        wf.wrapper_path("gatk3_ug")
 
 
 # Run GATK 4 HaplotypeCaller Joint (direct, pedigree/multi-sample) ------------
@@ -174,28 +190,6 @@ rule variant_calling_gatk4_hc_gvcf_genotype:
         wf.wrapper_path("gatk4_hc/genotype")
 
 
-# Run GATK UnifiedGenotyper ---------------------------------------------------
-
-
-rule variant_calling_gatk_ug_run:
-    input:
-        unpack(wf.get_input_files("gatk_ug", "run")),
-    output:
-        **wf.get_output_files("gatk_ug", "run"),
-    threads: wf.get_resource("gatk_ug", "run", "threads")
-    resources:
-        time=wf.get_resource("gatk_ug", "run", "time"),
-        memory=wf.get_resource("gatk_ug", "run", "memory"),
-        partition=wf.get_resource("gatk_ug", "run", "partition"),
-    params:
-        step_key="variant_calling",
-        caller_key="gatk_ug",
-    log:
-        **wf.get_log_file("gatk_ug", "run"),
-    wrapper:
-        wf.wrapper_path("gatk_ug_par")
-
-
 # QC / Statistics ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 # Generate BCFtools stats report ----------------------------------------------

snappy_pipeline/workflows/variant_calling/__init__.py

@@ -146,7 +146,6 @@
 from snappy_pipeline.workflows.abstract import (
     BaseStep,
     BaseStepPart,
-    LinkOutStepPart,
     ResourceUsage,
     WritePedigreeStepPart,
 )
@@ -204,78 +203,20 @@
     gatk4_hc_joint:
       window_length: 10000000
       num_threads: 16
-      allow_seq_dict_incompatibility: false  # REQUIRED
-      annotations: []  # REQUIRED
+      allow_seq_dict_incompatibility: false
     gatk4_hc_gvcf:
       window_length: 10000000
       num_threads: 16
-      allow_seq_dict_incompatibility: false  # REQUIRED
-      annotations: []  # REQUIRED
-    gatk_hc:
-      # Parallelization configuration
-      num_threads: 2            # number of cores to use locally
-      window_length: 5000000    # split input into windows of this size, each triggers a job
-      num_jobs: 500             # number of windows to process in parallel
-      use_profile: true         # use Snakemake profile for parallel processing
-      restart_times: 0          # number of times to re-launch jobs in case of failure
-      max_jobs_per_second: 10   # throttling of job creation
-      max_status_checks_per_second: 10  # throttling of status jobs
-      debug_trunc_tokens: 0     # truncation to first N tokens (0 for none)
-      keep_tmpdir: never        # keep temporary directory, {always, never, onerror}
-      job_mult_memory: 1        # memory multiplier
-      job_mult_time: 1          # running time multiplier
-      merge_mult_memory: 1      # memory multiplier for merging
-      merge_mult_time: 1        # running time multiplier for merging
-      # GATK HC--specific configuration
       allow_seq_dict_incompatibility: false
-      annotations:
-      - BaseQualityRankSumTest
-      - FisherStrand
-      - GCContent
-      - HaplotypeScore
-      - HomopolymerRun
-      - MappingQualityRankSumTest
-      - MappingQualityZero
-      - QualByDepth
-      - ReadPosRankSumTest
-      - RMSMappingQuality
-      - DepthPerAlleleBySample
-      - Coverage
-      - ClippingRankSumTest
-      - DepthPerSampleHC
-    gatk_ug:
-      # Parallelization configuration
-      num_threads: 2            # number of cores to use locally
-      window_length: 5000000    # split input into windows of this size, each triggers a job
-      num_jobs: 500             # number of windows to process in parallel
-      use_profile: true         # use Snakemake profile for parallel processing
-      restart_times: 0          # number of times to re-launch jobs in case of failure
-      max_jobs_per_second: 10   # throttling of job creation
-      max_status_checks_per_second: 10  # throttling of status jobs
-      debug_trunc_tokens: 0     # truncation to first N tokens (0 for none)
-      keep_tmpdir: never        # keep temporary directory, {always, never, onerror}
-      job_mult_memory: 1        # memory multiplier
-      job_mult_time: 1          # running time multiplier
-      merge_mult_memory: 1      # memory multiplier for merging
-      merge_mult_time: 1        # running time multiplier for merging
-      # GATK UG--specific configuration
+    gatk3_hc:
+      num_threads: 16
+      window_length: 10000000
+      allow_seq_dict_incompatibility: false
+    gatk3_ug:
+      num_threads: 16
+      window_length: 10000000
       allow_seq_dict_incompatibility: false
       downsample_to_coverage: 250
-      annotations:
-      - BaseQualityRankSumTest
-      - FisherStrand
-      - GCContent
-      - HaplotypeScore
-      - HomopolymerRun
-      - MappingQualityRankSumTest
-      - MappingQualityZero
-      - QualByDepth
-      - ReadPosRankSumTest
-      - RMSMappingQuality
-      - DepthPerAlleleBySample
-      - Coverage
-      - ClippingRankSumTest
-      - DepthPerSampleHC
 """
 
 
@@ -467,8 +408,8 @@ def get_resource_usage(self, action):
         )
 
 
-class GatkCallerStepPartBase(VariantCallingStepPart):
-    """Germlin variant calling with GATK caller"""
+class Gatk3CallerStepPartBase(VariantCallingStepPart):
+    """Germlin variant calling with GATK v3 caller"""
 
     def check_config(self):
         if self.__class__.name not in self.config["tools"]:
@@ -495,24 +436,24 @@ def get_resource_usage(self, action):
         )
 
 
-class GatkHaplotypeCallerStepPart(GatkCallerStepPartBase):
+class Gatk3HaplotypeCallerStepPart(Gatk3CallerStepPartBase):
     """Germline variant calling with GATK HaplotypeCaller
 
-    This triggers the cluster-parallel variant calling with gatk_ug, GATK3.
+    This triggers the cluster-parallel variant calling with gatk3_hc, GATK3.
     """
 
     #: Step name
-    name = "gatk_hc"
+    name = "gatk3_hc"
 
 
-class GatkUnifiedGenotyperStepPart(GatkCallerStepPartBase):
+class Gatk3UnifiedGenotyperStepPart(Gatk3CallerStepPartBase):
     """Germline variant calling with GATK UnifiedGenotyper
 
-    This triggers the cluster-parallel variant calling with gatk_ug, GATK3.
+    This triggers the cluster-parallel variant calling with gatk3_ug, GATK3.
     """
 
     #: Step name
-    name = "gatk_ug"
+    name = "gatk3_ug"
 
 
 class Gatk4CallerStepPartBase(VariantCallingStepPart):
@@ -900,14 +841,13 @@ def __init__(self, workflow, config, config_lookup_paths, config_paths, workdir)
             (
                 WritePedigreeStepPart,
                 BcftoolsStepPart,
-                GatkHaplotypeCallerStepPart,
-                GatkUnifiedGenotyperStepPart,
+                Gatk3HaplotypeCallerStepPart,
+                Gatk3UnifiedGenotyperStepPart,
                 Gatk4HaplotypeCallerJointStepPart,
                 Gatk4HaplotypeCallerGvcfStepPart,
                 BcftoolsStatsStepPart,
                 JannovarStatisticsStepPart,
                 BafFileGenerationStepPart,
-                LinkOutStepPart,
             )
         )
         # Register sub workflows

snappy_wrappers/wrappers/gatk3_hc/environment.yaml

@@ -0,0 +1 @@
+../gatk3_ug/environment.yaml