Merge pull request #87 from uclahs-cds/nwiltsie-update-reference-paths

Update cluster reference paths
uclahs-cds · Oct 25, 2024 · d40c8b0 · d40c8b0
2 parents 7aaae34 + b406b37
commit d40c8b0
Show file tree

Hide file tree

Showing 6 changed files with 44 additions and 44 deletions.
diff --git a/README.md b/README.md
@@ -144,11 +144,11 @@ For normal-only or tumour-only samples, exclude the fields for the other state.
 | `scatter_count` | Yes | integer | Number of intervals to divide into for parallelization |
 | `intervals` | Yes | path | Use all .list in inputs for WGS; Set to absolute path to targeted exome interval file (with .interval_list, .list, .intervals, or .bed suffix) |
 | `gatk_ir_compression` | No | integer | Compression level for BAMs output by IndelRealigner. Default: 0. Range: 0-9 |
-| `reference_fasta` | Yes | path | Absolute path to reference genome fasta file, e.g., `/hot/ref/reference/GRCh38-BI-20160721/Homo_sapiens_assembly38.fasta` |
-| `bundle_mills_and_1000g_gold_standard_indels_vcf_gz` | Yes | path | Absolute path to Mills & 1000G Gold Standard Indels file, e.g., `/hot/ref/tool-specific-input/GATK/GRCh38/Mills_and_1000G_gold_standard.indels.hg38.vcf.gz` |
-| `bundle_known_indels_vcf_gz` | Yes | path | Absolute path to known indels file, e.g., `/hot/ref/tool-specific-input/GATK/GRCh38/Homo_sapiens_assembly38.known_indels.vcf.gz` |
-| `bundle_v0_dbsnp138_vcf_gz` | Yes | path | Absolute path to dbsnp file, e.g., `/hot/ref/tool-specific-input/GATK/GRCh38/resources_broad_hg38_v0_Homo_sapiens_assembly38.dbsnp138.vcf.gz` |
-| `bundle_contest_hapmap_3p3_vcf_gz` | Yes | path | Absolute path to HapMap 3.3 biallelic sites file, e.g., `/hot/ref/tool-specific-input/GATK/GRCh38/Biallelic/hapmap_3.3.hg38.BIALLELIC.PASS.2021-09-01.vcf.gz` |
+| `reference_fasta` | Yes | path | Absolute path to reference genome fasta file, e.g., `/hot/resource/reference-genome/GRCh38-BI-20160721/Homo_sapiens_assembly38.fasta` |
+| `bundle_mills_and_1000g_gold_standard_indels_vcf_gz` | Yes | path | Absolute path to Mills & 1000G Gold Standard Indels file, e.g., `/hot/resource/tool-specific-input/GATK/GRCh38/Mills_and_1000G_gold_standard.indels.hg38.vcf.gz` |
+| `bundle_known_indels_vcf_gz` | Yes | path | Absolute path to known indels file, e.g., `/hot/resource/tool-specific-input/GATK/GRCh38/Homo_sapiens_assembly38.known_indels.vcf.gz` |
+| `bundle_v0_dbsnp138_vcf_gz` | Yes | path | Absolute path to dbsnp file, e.g., `/hot/resource/tool-specific-input/GATK/GRCh38/resources_broad_hg38_v0_Homo_sapiens_assembly38.dbsnp138.vcf.gz` |
+| `bundle_contest_hapmap_3p3_vcf_gz` | Yes | path | Absolute path to HapMap 3.3 biallelic sites file, e.g., `/hot/resource/tool-specific-input/GATK/GRCh38/Biallelic/hapmap_3.3.hg38.BIALLELIC.PASS.2021-09-01.vcf.gz` |
 | `work_dir` | optional | path | Path of working directory for Nextflow. When included in the sample config file, Nextflow intermediate files and logs will be saved to this directory. With ucla_cds, the default is `/scratch` and should only be changed for testing/development. Changing this directory to `/hot` or `/tmp` can lead to high server latency and potential disk space limitations, respectively. |
 | `base_resource_update` | optional | namespace | Namespace of parameters to update base resource allocations in the pipeline. Usage and structure are detailed in `template.config` and below. |
 

diff --git a/config/template.config b/config/template.config
@@ -41,8 +41,8 @@ params {
 
     // Reference - Used here hg38 decoy version
     // GATK requires the reference fasta to be accompanied by a .fai index and .dict dictionary associated with the fasta for fast random access
-    // These can be found in the same folder as the reference here: /hot/ref/reference/GRCh38-BI-20160721
-    reference_fasta = "/hot/ref/reference/GRCh38-BI-20160721/Homo_sapiens_assembly38.fasta"
+    // These can be found in the same folder as the reference here: /hot/resource/reference-genome/GRCh38-BI-20160721
+    reference_fasta = "/hot/resource/reference-genome/GRCh38-BI-20160721/Homo_sapiens_assembly38.fasta"
 
     // Whether to parallelize the pipeline by chromosome or by splitting into equal-sized intervals
     // The scatter_count and extra args below only go into effect if parallelize_by_chromosome is disabled
@@ -51,10 +51,10 @@ params {
     split_intervals_extra_args = ''
 
     // GATK bundle - Used here hg38 decoy version
-    bundle_mills_and_1000g_gold_standard_indels_vcf_gz = "/hot/ref/tool-specific-input/GATK/GRCh38/Mills_and_1000G_gold_standard.indels.hg38.vcf.gz"
-    bundle_known_indels_vcf_gz = "/hot/ref/tool-specific-input/GATK/GRCh38/Homo_sapiens_assembly38.known_indels.vcf.gz"
-    bundle_v0_dbsnp138_vcf_gz = "/hot/ref/tool-specific-input/GATK/GRCh38/resources_broad_hg38_v0_Homo_sapiens_assembly38.dbsnp138.vcf.gz"
-    bundle_contest_hapmap_3p3_vcf_gz = "/hot/ref/tool-specific-input/GATK/GRCh38/Biallelic/hapmap_3.3.hg38.BIALLELIC.PASS.2021-09-01.vcf.gz"
+    bundle_mills_and_1000g_gold_standard_indels_vcf_gz = "/hot/resource/tool-specific-input/GATK/GRCh38/Mills_and_1000G_gold_standard.indels.hg38.vcf.gz"
+    bundle_known_indels_vcf_gz = "/hot/resource/tool-specific-input/GATK/GRCh38/Homo_sapiens_assembly38.known_indels.vcf.gz"
+    bundle_v0_dbsnp138_vcf_gz = "/hot/resource/tool-specific-input/GATK/GRCh38/resources_broad_hg38_v0_Homo_sapiens_assembly38.dbsnp138.vcf.gz"
+    bundle_contest_hapmap_3p3_vcf_gz = "/hot/resource/tool-specific-input/GATK/GRCh38/Biallelic/hapmap_3.3.hg38.BIALLELIC.PASS.2021-09-01.vcf.gz"
 
     // Base resource allocation updater
     // See README for adding parameters to update the base resource allocations

diff --git a/test/configtest-F16.json b/test/configtest-F16.json
@@ -52,14 +52,14 @@
     "params": {
       "aligner": "BWA-MEM2-2.2.1",
       "blcds_registered_dataset": false,
-      "bundle_contest_hapmap_3p3_vcf_gz": "/hot/ref/tool-specific-input/GATK/GRCh38/Biallelic/hapmap_3.3.hg38.BIALLELIC.PASS.2021-09-01.vcf.gz",
-      "bundle_contest_hapmap_3p3_vcf_gz_tbi": "/hot/ref/tool-specific-input/GATK/GRCh38/Biallelic/hapmap_3.3.hg38.BIALLELIC.PASS.2021-09-01.vcf.gz.tbi",
-      "bundle_known_indels_vcf_gz": "/hot/ref/tool-specific-input/GATK/GRCh38/Homo_sapiens_assembly38.known_indels.vcf.gz",
-      "bundle_known_indels_vcf_gz_tbi": "/hot/ref/tool-specific-input/GATK/GRCh38/Homo_sapiens_assembly38.known_indels.vcf.gz.tbi",
-      "bundle_mills_and_1000g_gold_standard_indels_vcf_gz": "/hot/ref/tool-specific-input/GATK/GRCh38/Mills_and_1000G_gold_standard.indels.hg38.vcf.gz",
-      "bundle_mills_and_1000g_gold_standard_indels_vcf_gz_tbi": "/hot/ref/tool-specific-input/GATK/GRCh38/Mills_and_1000G_gold_standard.indels.hg38.vcf.gz.tbi",
-      "bundle_v0_dbsnp138_vcf_gz": "/hot/ref/tool-specific-input/GATK/GRCh38/resources_broad_hg38_v0_Homo_sapiens_assembly38.dbsnp138.vcf.gz",
-      "bundle_v0_dbsnp138_vcf_gz_tbi": "/hot/ref/tool-specific-input/GATK/GRCh38/resources_broad_hg38_v0_Homo_sapiens_assembly38.dbsnp138.vcf.gz.tbi",
+      "bundle_contest_hapmap_3p3_vcf_gz": "/hot/resource/tool-specific-input/GATK/GRCh38/Biallelic/hapmap_3.3.hg38.BIALLELIC.PASS.2021-09-01.vcf.gz",
+      "bundle_contest_hapmap_3p3_vcf_gz_tbi": "/hot/resource/tool-specific-input/GATK/GRCh38/Biallelic/hapmap_3.3.hg38.BIALLELIC.PASS.2021-09-01.vcf.gz.tbi",
+      "bundle_known_indels_vcf_gz": "/hot/resource/tool-specific-input/GATK/GRCh38/Homo_sapiens_assembly38.known_indels.vcf.gz",
+      "bundle_known_indels_vcf_gz_tbi": "/hot/resource/tool-specific-input/GATK/GRCh38/Homo_sapiens_assembly38.known_indels.vcf.gz.tbi",
+      "bundle_mills_and_1000g_gold_standard_indels_vcf_gz": "/hot/resource/tool-specific-input/GATK/GRCh38/Mills_and_1000G_gold_standard.indels.hg38.vcf.gz",
+      "bundle_mills_and_1000g_gold_standard_indels_vcf_gz_tbi": "/hot/resource/tool-specific-input/GATK/GRCh38/Mills_and_1000G_gold_standard.indels.hg38.vcf.gz.tbi",
+      "bundle_v0_dbsnp138_vcf_gz": "/hot/resource/tool-specific-input/GATK/GRCh38/resources_broad_hg38_v0_Homo_sapiens_assembly38.dbsnp138.vcf.gz",
+      "bundle_v0_dbsnp138_vcf_gz_tbi": "/hot/resource/tool-specific-input/GATK/GRCh38/resources_broad_hg38_v0_Homo_sapiens_assembly38.dbsnp138.vcf.gz.tbi",
       "cache_intermediate_pipeline_steps": false,
       "dataset_id": "A-mini",
       "docker_container_registry": "ghcr.io/uclahs-cds",
@@ -75,7 +75,7 @@
       "input": {
         "BAM": {
           "tumor": [
-            "/hot/resource/SMC-HET/tumours/A-mini/bams/n1/output/S2.T-n1.bam"
+            "/hot/data/unregistered/SMC-HET/tumours/A-mini/bams/n1/output/S2.T-n1.bam"
           ]
         },
         "recalibration_table": [
@@ -228,13 +228,13 @@
           "memory": "1 GB"
         }
       },
-      "reference_fasta": "/hot/ref/reference/GRCh38-BI-20160721/Homo_sapiens_assembly38.fasta",
-      "reference_fasta_dict": "/hot/ref/reference/GRCh38-BI-20160721/Homo_sapiens_assembly38.dict",
-      "reference_fasta_fai": "/hot/ref/reference/GRCh38-BI-20160721/Homo_sapiens_assembly38.fasta.fai",
+      "reference_fasta": "/hot/resource/reference-genome/GRCh38-BI-20160721/Homo_sapiens_assembly38.fasta",
+      "reference_fasta_dict": "/hot/resource/reference-genome/GRCh38-BI-20160721/Homo_sapiens_assembly38.dict",
+      "reference_fasta_fai": "/hot/resource/reference-genome/GRCh38-BI-20160721/Homo_sapiens_assembly38.fasta.fai",
       "samples_to_process": [
         {
           "id": "4915723",
-          "path": "/hot/resource/SMC-HET/tumours/A-mini/bams/n1/output/S2.T-n1.bam",
+          "path": "/hot/data/unregistered/SMC-HET/tumours/A-mini/bams/n1/output/S2.T-n1.bam",
           "sample_type": "tumor"
         }
       ],

diff --git a/test/configtest-F32.json b/test/configtest-F32.json
@@ -52,14 +52,14 @@
     "params": {
       "aligner": "BWA-MEM2-2.2.1",
       "blcds_registered_dataset": false,
-      "bundle_contest_hapmap_3p3_vcf_gz": "/hot/ref/tool-specific-input/GATK/GRCh38/Biallelic/hapmap_3.3.hg38.BIALLELIC.PASS.2021-09-01.vcf.gz",
-      "bundle_contest_hapmap_3p3_vcf_gz_tbi": "/hot/ref/tool-specific-input/GATK/GRCh38/Biallelic/hapmap_3.3.hg38.BIALLELIC.PASS.2021-09-01.vcf.gz.tbi",
-      "bundle_known_indels_vcf_gz": "/hot/ref/tool-specific-input/GATK/GRCh38/Homo_sapiens_assembly38.known_indels.vcf.gz",
-      "bundle_known_indels_vcf_gz_tbi": "/hot/ref/tool-specific-input/GATK/GRCh38/Homo_sapiens_assembly38.known_indels.vcf.gz.tbi",
-      "bundle_mills_and_1000g_gold_standard_indels_vcf_gz": "/hot/ref/tool-specific-input/GATK/GRCh38/Mills_and_1000G_gold_standard.indels.hg38.vcf.gz",
-      "bundle_mills_and_1000g_gold_standard_indels_vcf_gz_tbi": "/hot/ref/tool-specific-input/GATK/GRCh38/Mills_and_1000G_gold_standard.indels.hg38.vcf.gz.tbi",
-      "bundle_v0_dbsnp138_vcf_gz": "/hot/ref/tool-specific-input/GATK/GRCh38/resources_broad_hg38_v0_Homo_sapiens_assembly38.dbsnp138.vcf.gz",
-      "bundle_v0_dbsnp138_vcf_gz_tbi": "/hot/ref/tool-specific-input/GATK/GRCh38/resources_broad_hg38_v0_Homo_sapiens_assembly38.dbsnp138.vcf.gz.tbi",
+      "bundle_contest_hapmap_3p3_vcf_gz": "/hot/resource/tool-specific-input/GATK/GRCh38/Biallelic/hapmap_3.3.hg38.BIALLELIC.PASS.2021-09-01.vcf.gz",
+      "bundle_contest_hapmap_3p3_vcf_gz_tbi": "/hot/resource/tool-specific-input/GATK/GRCh38/Biallelic/hapmap_3.3.hg38.BIALLELIC.PASS.2021-09-01.vcf.gz.tbi",
+      "bundle_known_indels_vcf_gz": "/hot/resource/tool-specific-input/GATK/GRCh38/Homo_sapiens_assembly38.known_indels.vcf.gz",
+      "bundle_known_indels_vcf_gz_tbi": "/hot/resource/tool-specific-input/GATK/GRCh38/Homo_sapiens_assembly38.known_indels.vcf.gz.tbi",
+      "bundle_mills_and_1000g_gold_standard_indels_vcf_gz": "/hot/resource/tool-specific-input/GATK/GRCh38/Mills_and_1000G_gold_standard.indels.hg38.vcf.gz",
+      "bundle_mills_and_1000g_gold_standard_indels_vcf_gz_tbi": "/hot/resource/tool-specific-input/GATK/GRCh38/Mills_and_1000G_gold_standard.indels.hg38.vcf.gz.tbi",
+      "bundle_v0_dbsnp138_vcf_gz": "/hot/resource/tool-specific-input/GATK/GRCh38/resources_broad_hg38_v0_Homo_sapiens_assembly38.dbsnp138.vcf.gz",
+      "bundle_v0_dbsnp138_vcf_gz_tbi": "/hot/resource/tool-specific-input/GATK/GRCh38/resources_broad_hg38_v0_Homo_sapiens_assembly38.dbsnp138.vcf.gz.tbi",
       "cache_intermediate_pipeline_steps": false,
       "dataset_id": "A-mini",
       "docker_container_registry": "ghcr.io/uclahs-cds",
@@ -75,7 +75,7 @@
       "input": {
         "BAM": {
           "tumor": [
-            "/hot/resource/SMC-HET/tumours/A-mini/bams/n1/output/S2.T-n1.bam"
+            "/hot/data/unregistered/SMC-HET/tumours/A-mini/bams/n1/output/S2.T-n1.bam"
           ]
         },
         "recalibration_table": [
@@ -228,13 +228,13 @@
           "memory": "1 GB"
         }
       },
-      "reference_fasta": "/hot/ref/reference/GRCh38-BI-20160721/Homo_sapiens_assembly38.fasta",
-      "reference_fasta_dict": "/hot/ref/reference/GRCh38-BI-20160721/Homo_sapiens_assembly38.dict",
-      "reference_fasta_fai": "/hot/ref/reference/GRCh38-BI-20160721/Homo_sapiens_assembly38.fasta.fai",
+      "reference_fasta": "/hot/resource/reference-genome/GRCh38-BI-20160721/Homo_sapiens_assembly38.fasta",
+      "reference_fasta_dict": "/hot/resource/reference-genome/GRCh38-BI-20160721/Homo_sapiens_assembly38.dict",
+      "reference_fasta_fai": "/hot/resource/reference-genome/GRCh38-BI-20160721/Homo_sapiens_assembly38.fasta.fai",
       "samples_to_process": [
         {
           "id": "4915723",
-          "path": "/hot/resource/SMC-HET/tumours/A-mini/bams/n1/output/S2.T-n1.bam",
+          "path": "/hot/data/unregistered/SMC-HET/tumours/A-mini/bams/n1/output/S2.T-n1.bam",
           "sample_type": "tumor"
         }
       ],

diff --git a/test/nftest.config b/test/nftest.config
@@ -39,8 +39,8 @@ params {
 
     // Reference - Used here hg38 decoy version
     // GATK requires the reference fasta to be accompanied by a .fai index and .dict dictionary associated with the fasta for fast random access
-    // These can be found in the same folder as the reference here: /hot/ref/reference/GRCh38-BI-20160721
-    reference_fasta = "/hot/ref/reference/GRCh38-BI-20160721/Homo_sapiens_assembly38.fasta"
+    // These can be found in the same folder as the reference here: /hot/resource/reference-genome/GRCh38-BI-20160721
+    reference_fasta = "/hot/resource/reference-genome/GRCh38-BI-20160721/Homo_sapiens_assembly38.fasta"
 
     // Whether to parallelize the pipeline by chromosome or by splitting into equal-sized intervals
     // The scatter_count and extra args below only go into effect if parallelize_by_chromosome is disabled
@@ -49,10 +49,10 @@ params {
     split_intervals_extra_args = ''
 
     // GATK bundle - Used here hg38 decoy version
-    bundle_mills_and_1000g_gold_standard_indels_vcf_gz = "/hot/ref/tool-specific-input/GATK/GRCh38/Mills_and_1000G_gold_standard.indels.hg38.vcf.gz"
-    bundle_known_indels_vcf_gz = "/hot/ref/tool-specific-input/GATK/GRCh38/Homo_sapiens_assembly38.known_indels.vcf.gz"
-    bundle_v0_dbsnp138_vcf_gz = "/hot/ref/tool-specific-input/GATK/GRCh38/resources_broad_hg38_v0_Homo_sapiens_assembly38.dbsnp138.vcf.gz"
-    bundle_contest_hapmap_3p3_vcf_gz = "/hot/ref/tool-specific-input/GATK/GRCh38/Biallelic/hapmap_3.3.hg38.BIALLELIC.PASS.2021-09-01.vcf.gz"
+    bundle_mills_and_1000g_gold_standard_indels_vcf_gz = "/hot/resource/tool-specific-input/GATK/GRCh38/Mills_and_1000G_gold_standard.indels.hg38.vcf.gz"
+    bundle_known_indels_vcf_gz = "/hot/resource/tool-specific-input/GATK/GRCh38/Homo_sapiens_assembly38.known_indels.vcf.gz"
+    bundle_v0_dbsnp138_vcf_gz = "/hot/resource/tool-specific-input/GATK/GRCh38/resources_broad_hg38_v0_Homo_sapiens_assembly38.dbsnp138.vcf.gz"
+    bundle_contest_hapmap_3p3_vcf_gz = "/hot/resource/tool-specific-input/GATK/GRCh38/Biallelic/hapmap_3.3.hg38.BIALLELIC.PASS.2021-09-01.vcf.gz"
 }
 
 // Setup the pipeline config. DO NOT REMOVE THIS LINE!

diff --git a/test/single.yaml b/test/single.yaml
@@ -3,4 +3,4 @@ patient_id: TWGSAMIN000001
 input:
   BAM:
     tumor:
-      - "/hot/resource/SMC-HET/tumours/A-mini/bams/n1/output/S2.T-n1.bam"
+      - "/hot/data/unregistered/SMC-HET/tumours/A-mini/bams/n1/output/S2.T-n1.bam"
-Original file line number
+Diff line change
@@ Expand Up / @@ -3,4 +3,4 @@ patient_id: TWGSAMIN000001 @@
     input:
       BAM:
         tumor:
-          - "/hot/resource/SMC-HET/tumours/A-mini/bams/n1/output/S2.T-n1.bam"
+          - "/hot/data/unregistered/SMC-HET/tumours/A-mini/bams/n1/output/S2.T-n1.bam"