update schema

uclahs-cds · Oct 24, 2024 · 3d8167a · 3d8167a
1 parent f88ab7f
commit 3d8167a
Show file tree

Hide file tree

Showing 4 changed files with 94 additions and 70 deletions.
diff --git a/README.md b/README.md
@@ -99,8 +99,8 @@ input:
 | `liftover_direction`        | string | Conversion direction: [GRCh37ToGRCh38, GRCh38ToGRCh37].                                                                                            |
 | `fasta_ref_37`              | path   | Path to the GRCh37 reference sequence (FASTA).                                                                                                     |
 | `fasta_ref_38`              | path   | Path to the GRCh38 reference sequence (FASTA).                                                                                                     |
-| `funcotator_data_source`    | path   | Path to [Funcotator data source](https://gatk.broadinstitute.org/hc/en-us/articles/360050815792-FuncotatorDataSourceDownloader) directory containing dbSNP, GENCODE and HGNC sources for SNV annotation.         |
 | `resource_bundle_path`      | path   | Path to unpacked [resource-bundle.zip](https://github.com/uclahs-cds/pipeline-StableLift/releases/download/v1.1.0/resource-bundle.zip).            |
+| `funcotator_data_source`    | path   | Path to [Funcotator data source](https://gatk.broadinstitute.org/hc/en-us/articles/360050815792-FuncotatorDataSourceDownloader) directory containing dbSNP, GENCODE and HGNC sources (required for SNV annotation).|
 
 | Optional Parameter          | Type                                                                                      | Default                      | Description                                                                                                                                                                                                                                                                                                                                                                           |
 | --------------------------- | ----------------------------------------------------------------------------------------- | ---------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |

diff --git a/config/schema-snv.yaml b/config/schema-snv.yaml
@@ -3,10 +3,10 @@ repeat_bed:
   type: 'Path'
   mode: 'r'
   required: true
-  help: 'RepeatMasker (v3.0.1) intervals from UCSC Table Browser for variant annotation in GRCh38 coordinates, included in resource-bundle.zip'
+  help: 'RepeatMasker (v3.0.1) intervals from UCSC Table Browser for variant annotation in target build coordinates, included in resource-bundle.zip'
 
 funcotator_data_source:
   type: 'Path'
-  required: true
   mode: 'r'
-  help: 'Root data source folder for Funcotator from https://gatk.broadinstitute.org/hc/en-us/articles/360035889931-Funcotator-Information-and-Tutorial'
+  required: true
+  help: 'Path to Funcotator data source directory containing dbSNP, GENCODE and HGNC sources for SNV annotation'
diff --git a/config/schema.yaml b/config/schema.yaml
@@ -1,53 +1,105 @@
 ---
-sample_id:
-  type: 'String'
-  required: true
-  help: 'Sample id supplied from input yaml'
-
-liftover_direction:
-  type: 'String'
+# Required Parameters
+output_dir:
+  type: 'Path'
+  mode: 'w'
   required: true
-  help: 'Direction of LiftOver to perform'
-  choices:
-    - GRCh37ToGRCh38
-    - GRCh38ToGRCh37
+  help: 'Path to the directory where the output files are to be saved'
 
 variant_caller:
   type: 'String'
   required: true
-  help: 'Tool used to call variants'
+  help: 'Variant calling algorithm used to generate input VCF'
   choices:
-    - Mutect2
     - HaplotypeCaller
+    - Mutect2
     - Strelka2
-    - Muse2
     - SomaticSniper
+    - Muse2
     - Delly2-gSV
     - Delly2-sSV
 
-save_intermediate_files:
-  type: 'Bool'
+rf_model:
+  type: 'Path'
+  mode: 'r'
   required: true
-  default: false
-  help: 'Enable to store intermediate files'
+  help: 'Path to corresponding pre-trained random forest model'
 
-output_dir:
-  type: 'Path'
-  mode: 'w'
+liftover_direction:
+  type: 'String'
   required: true
-  help: 'Absolute path to directory to store output'
+  help: 'Conversion direction'
+  choices:
+    - GRCh37ToGRCh38
+    - GRCh38ToGRCh37
 
 fasta_ref_37:
   type: 'Path'
   mode: 'r'
   required: true
-  help: 'GRCh37 FASTA reference'
+  help: 'Path to the GRCh37 reference sequence (FASTA)'
 
 fasta_ref_38:
   type: 'Path'
   mode: 'r'
   required: true
-  help: 'GRCh38 FASTA reference'
+  help: 'Path to the GRCh38 reference sequence (FASTA)'
+
+resource_bundle_path:
+  type: 'Path'
+  mode: 'r'
+  required: true
+  help: 'Path to unpacked resource-bundle.zip'
+
+# Optional Parameters
+target_threshold:
+  type: 'RangedNumber'
+  required: false
+  min: 0
+  max: 1
+  default: ''
+  help: 'Target Stability Score threshold for variant filtering'
+
+target_specificity:
+  type: 'RangedNumber'
+  required: false
+  min: 0
+  max: 1
+  default: ''
+  help: 'Target specificity based on whole genome validation set for variant filtering'
+
+extract_features_cpus:
+  type: 'Integer'
+  required: false
+  default: 4
+  help: 'Number of cpus to use for parallel parsing of large VCFs (>1GB)'
+
+save_intermediate_files:
+  type: 'Bool'
+  required: false
+  default: false
+  help: 'If set, save output files from intermediate pipeline processes'
+
+blcds_registered_dataset:
+  type: 'Bool'
+  required: false
+  default: false
+  help: 'Set to true when using BLCDS folder structure; use false for now'
+
+ucla_cds:
+  type: 'Bool'
+  required: false
+  default: false
+  help: 'If set, overwrite default memory and CPU values by UCLA cluster-specific configs'
+
+# Internal parameters
+src_fasta_id:
+  type: 'String'
+  required: true
+  help: 'Source reference genome build identifier'
+  choices:
+    - GRCh37
+    - GRCh38
 
 src_fasta_ref:
   type: 'Path'
@@ -67,6 +119,14 @@ src_fasta_dict:
   required: true
   help: 'Source reference sequence dictionary'
 
+dest_fasta_id:
+  type: 'String'
+  required: true
+  help: 'Destination reference genome build identifier'
+  choices:
+    - GRCh37
+    - GRCh38
+
 dest_fasta_ref:
   type: 'Path'
   mode: 'r'
@@ -89,41 +149,4 @@ chain_file:
   type: 'Path'
   mode: 'r'
   required: true
-  help: 'Chain file corresponding to LiftOver direction, included in resource-bundle.zip'
-
-rf_model:
-  type: 'Path'
-  mode: 'r'
-  required: true
-  help: 'Path to pre-trained random forest model (.Rds) corresponding to variant caller and LiftOver direction'
-
-input:
-  type: 'Namespace'
-  required: true
-  help: 'Input sample'
-  elements:
-    vcf:
-      type: 'Path'
-      mode: 'r'
-      required: true
-      help: 'Input dataset supplied by input yaml'
-
-target_threshold:
-  type: 'RangedNumber'
-  required: false
-  min: 0
-  max: 1
-  help: >-
-    Optional parameter specifying target Stability Score threshold for variant
-    filtering Default behavior without `target_threshold` or
-    `target_specificity` specified uses threshold maximizing F1-score in whole
-    genome validation set'.
-
-target_specificity:
-  type: 'RangedNumber'
-  required: false
-  min: 0
-  max: 1
-  help: >-
-    Optional parameter specifying target specificity for variant filtering
-    based on whole genome validation set. Overrides `target_threshold`.
+  help: 'Chain file for LiftOver conversion'
diff --git a/config/template.config b/config/template.config
@@ -19,16 +19,17 @@ params {
     // Path to pre-trained random forest model
     rf_model = ""
 
-    // Path to reference fasta files
+    // Path to reference fasta files with corresponding index file (.fai) and sequence dictionary (.dict)
     fasta_ref_37 = "" // GRCh37-EBI-hs37d5/hs37d5.fa
     fasta_ref_38 = "" // GRCh38-BI-20160721/Homo_sapiens_assembly38.fasta
 
-    // Path to Funcotator data source directory containing dbSNP, GENCODE and HGNC sources for SNV annotation
-    funcotator_data_source = ""
-
     // Path to unpacked resource-bundle.zip
     resource_bundle_path = ""
 
+    // Path to Funcotator data source directory containing dbSNP, GENCODE and HGNC sources (required for SNV annotation)
+    // https://gatk.broadinstitute.org/hc/en-us/articles/360050815792-FuncotatorDataSourceDownloader
+    funcotator_data_source = ""
+
     // Optional parameter specifying target Stability Score threshold for variant filtering
     // Default behavior without `target_threshold` or `target_specificity` uses threshold maximizing F1-score in whole genome validation set
     // Must be in the range [0.0, 1.0]