Merge pull request #9 from PacificBiosciences/gconcepcion/yak-changes

Gconcepcion/yak changes
PacificBiosciences · Oct 26, 2023 · de41388 · de41388
2 parents f1e76ea + 05312fe
commit de41388
Show file tree

Hide file tree

Showing 11 changed files with 316 additions and 122 deletions.
diff --git a/README.md b/README.md
@@ -204,7 +204,7 @@ These files are hosted publicly in each of the cloud backends; see `backends/${b
 | Type | Name | Description | Notes |
 | :- | :- | :- | :- |
 | String | name | Reference name; used to name outputs (e.g., "GRCh38") | |
-| [IndexData](https://github.com/PacificBiosciences/wdl-common/blob/main/wdl/structs.wdl) | fasta | Reference genome and index | |
+| Array[[IndexData](https://github.com/PacificBiosciences/wdl-common/blob/main/wdl/structs.wdl)] | fastas | Reference genomes and associatedindex | |
 
 ## Other inputs
 
@@ -229,9 +229,9 @@ These files will be output if `cohort.samples[sample]` is set to `true` for any
 | Array[Array[File]?] | assembly_noseq_gfas | Assembly graphs in [GFA format](https://github.com/chhylp123/hifiasm/blob/master/docs/source/interpreting-output.rst). | |
 | Array[Array[File]?] | assembly_lowQ_beds | Coordinates of low quality regions in BED format. | |
 | Array[Array[File]?] | assembly_stats | Assembly size and NG50 stats generated by [calN50](https://github.com/lh3/calN50). | |
-| Array[[IndexData](https://github.com/PacificBiosciences/wdl-common/blob/main/wdl/structs.wdl)?] | asm_bam | [minimap2](https://github.com/lh3/minimap2) alignment of assembly to reference. | |
-| Array[[IndexData](https://github.com/PacificBiosciences/wdl-common/blob/main/wdl/structs.wdl)?] | htsbox_vcf | Naive pileup variant calling of assembly against reference with [`htsbox`](https://github.com/lh3/htsbox) | |
-| Array[File?] | htsbox_vcf_stats | [`bcftools stats`](https://samtools.github.io/bcftools/bcftools.html#stats) summary statistics for `htsbox` variant calls | |
+| Array[Array[[IndexData](https://github.com/PacificBiosciences/wdl-common/blob/main/wdl/structs.wdl)?]] | asm_bam | [minimap2](https://github.com/lh3/minimap2) alignment of assembly to reference. | |
+| Array[Array[[IndexData](https://github.com/PacificBiosciences/wdl-common/blob/main/wdl/structs.wdl)?]] | htsbox_vcf | Naive pileup variant calling of assembly against reference with [`htsbox`](https://github.com/lh3/htsbox) | |
+| Array[Array[File?]] | htsbox_vcf_stats | [`bcftools stats`](https://samtools.github.io/bcftools/bcftools.html#stats) summary statistics for `htsbox` variant calls | |
 
 ## De novo assembly - trio
 
@@ -243,7 +243,7 @@ These files will be output if `cohort.de_novo_assembly_trio` is set to `true` an
 | Array[Array[File]]? | trio_assembly_noseq_gfas | Assembly graphs in [GFA format](https://github.com/chhylp123/hifiasm/blob/master/docs/source/interpreting-output.rst). | |
 | Array[Array[File]]? | trio_assembly_lowQ_beds | Coordinates of low quality regions in BED format. | |
 | Array[Array[File]]? | trio_assembly_stats | Assembly size and NG50 stats generated by [calN50](https://github.com/lh3/calN50). | |
-| Array[[IndexData](https://github.com/PacificBiosciences/wdl-common/blob/main/wdl/structs.wdl)]? | trio_asm_bams | [minimap2](https://github.com/lh3/minimap2) alignment of assembly to reference. | |
+| Array[Array[[IndexData](https://github.com/PacificBiosciences/wdl-common/blob/main/wdl/structs.wdl)]?] | trio_asm_bams | [minimap2](https://github.com/lh3/minimap2) alignment of assembly to reference. | |
 | Array[Map[String, String]]? | haplotype_key | Indication of which haplotype (`hap1`/`hap2`) corresponds to which parent. | |
 
 # Tool versions and Docker images

diff --git a/backends/aws/inputs.aws.json b/backends/aws/inputs.aws.json
@@ -13,13 +13,14 @@
     ],
     "run_de_novo_assembly_trio": "Boolean"
   },
-  "de_novo_assembly.reference": {
-    "name": "GRCh38",
+  "de_novo_assembly.references": [
+    {
+    "name": "String",
     "fasta": {
-      "data": "s3://dnastack-resources/dataset/GRCh38/human_GRCh38_no_alt_analysis_set.fasta",
-      "data_index": "s3://dnastack-resources/dataset/GRCh38/human_GRCh38_no_alt_analysis_set.fasta.fai"
+      "data": "File",
+      "data_index": "File"
     }
-  },
+  ],
   "de_novo_assembly.backend": "AWS",
   "de_novo_assembly.zones": "us-east-2a us-east-2b us-east-2c",
   "de_novo_assembly.aws_spot_queue_arn": "",

diff --git a/backends/azure/inputs.azure.json b/backends/azure/inputs.azure.json
@@ -13,13 +13,14 @@
     ],
     "run_de_novo_assembly_trio": "Boolean"
   },
-  "de_novo_assembly.reference": {
-    "name": "GRCh38",
+  "de_novo_assembly.references": [
+    {
+    "name": "String",
     "fasta": {
-      "data": "/datasetpbrarediseases/dataset/GRCh38/human_GRCh38_no_alt_analysis_set.fasta",
-      "data_index": "/datasetpbrarediseases/dataset/GRCh38/human_GRCh38_no_alt_analysis_set.fasta.fai"
+      "data": "File",
+      "data_index": "File"
     }
-  },
+  ],
   "de_novo_assembly.backend": "Azure",
   "de_novo_assembly.preemptible": "Boolean"
 }
diff --git a/backends/gcp/inputs.gcp.json b/backends/gcp/inputs.gcp.json
@@ -13,13 +13,14 @@
     ],
     "run_de_novo_assembly_trio": "Boolean"
   },
-  "de_novo_assembly.reference": {
-    "name": "GRCh38",
+  "de_novo_assembly.references": [
+    {
+    "name": "String",
     "fasta": {
-      "data": "gs://<pb_bucket>/dataset/GRCh38/human_GRCh38_no_alt_analysis_set.fasta",
-      "data_index": "gs://<pb_bucket>/dataset/GRCh38/human_GRCh38_no_alt_analysis_set.fasta.fai"
+      "data": "File",
+      "data_index": "File"
     }
-  },
+  ],
   "de_novo_assembly.backend": "GCP",
   "de_novo_assembly.zones": "String",
   "de_novo_assembly.preemptible": "Boolean"

diff --git a/backends/hpc/inputs.hpc.json b/backends/hpc/inputs.hpc.json
@@ -13,13 +13,14 @@
     ],
     "run_de_novo_assembly_trio": "Boolean"
   },
-  "de_novo_assembly.reference": {
-    "name": "GRCh38",
+  "de_novo_assembly.references": [
+    {
+    "name": "String",
     "fasta": {
-      "data": "<local_path_prefix>/dataset/GRCh38/human_GRCh38_no_alt_analysis_set.fasta",
-      "data_index": "<local_path_prefix>/dataset/GRCh38/human_GRCh38_no_alt_analysis_set.fasta.fai"
+      "data": "File",
+      "data_index": "File"
     }
-  },
+  ],
   "de_novo_assembly.backend": "HPC",
   "de_novo_assembly.preemptible": false
 }
diff --git a/wdl-ci.config.json b/wdl-ci.config.json
@@ -137,7 +137,7 @@
         },
         "yak_count": {
           "key": "yak_count",
-          "digest": "2ovi7jh4btl4sb7xr23ga6mxtd7nlq4s",
+          "digest": "6hlh6n3b3cqohtmjweg57of626he4c4v",
           "tests": [
             {
               "inputs": {
@@ -150,6 +150,8 @@
                   "${resources_file_path}/m64017_200108_232219.hifi_reads.fasta",
                   "${resources_file_path}/m64017_200112_090459.hifi_reads.fasta"
                 ],
+                "yak_params": "-b37",
+                "mem_gb": 70,
                 "runtime_attributes": "${default_runtime_attributes}"
               },
               "output_tests": {
@@ -173,7 +175,7 @@
       "tasks": {
         "hifiasm_assemble": {
           "key": "hifiasm_assemble",
-          "digest": "r4ikydzmdaed4hzsmc3t7efh6mz5e4mx",
+          "digest": "vhkzwee3f754jcjksog22uyps3j6myow",
           "tests": [
             {
               "inputs": {
@@ -230,7 +232,7 @@
         },
         "gfa2fa": {
           "key": "gfa2fa",
-          "digest": "liyb2m4cbkovxctcgaxwunqkn5az77ev",
+          "digest": "es7l5kyje3fiy5vxjnnsqg4fw6sitmdo",
           "tests": [
             {
               "inputs": {
@@ -262,7 +264,7 @@
         },
         "align_hifiasm": {
           "key": "align_hifiasm",
-          "digest": "77gs34t4c2i6epsg2epukfoaign2fmnt",
+          "digest": "4qf5jeepfn3jv3g2socql6xh7vmd4b7s",
           "tests": [
             {
               "inputs": {
@@ -325,6 +327,156 @@
       "name": "",
       "description": "",
       "tasks": {}
+    },
+    "workflows/wdl-common/wdl/tasks/glnexus.wdl": {
+      "key": "workflows/wdl-common/wdl/tasks/glnexus.wdl",
+      "name": "",
+      "description": "",
+      "tasks": {
+        "glnexus": {
+          "key": "glnexus",
+          "digest": "",
+          "tests": []
+        }
+      }
+    },
+    "workflows/wdl-common/wdl/tasks/mosdepth.wdl": {
+      "key": "workflows/wdl-common/wdl/tasks/mosdepth.wdl",
+      "name": "",
+      "description": "",
+      "tasks": {
+        "mosdepth": {
+          "key": "mosdepth",
+          "digest": "",
+          "tests": []
+        }
+      }
+    },
+    "workflows/wdl-common/wdl/tasks/pbsv_call.wdl": {
+      "key": "workflows/wdl-common/wdl/tasks/pbsv_call.wdl",
+      "name": "",
+      "description": "",
+      "tasks": {
+        "pbsv_call": {
+          "key": "pbsv_call",
+          "digest": "",
+          "tests": []
+        }
+      }
+    },
+    "workflows/wdl-common/wdl/tasks/pbsv_discover.wdl": {
+      "key": "workflows/wdl-common/wdl/tasks/pbsv_discover.wdl",
+      "name": "",
+      "description": "",
+      "tasks": {
+        "pbsv_discover": {
+          "key": "pbsv_discover",
+          "digest": "",
+          "tests": []
+        }
+      }
+    },
+    "workflows/wdl-common/wdl/tasks/pharmcat.wdl": {
+      "key": "workflows/wdl-common/wdl/tasks/pharmcat.wdl",
+      "name": "",
+      "description": "",
+      "tasks": {
+        "pangu_cyp2d6": {
+          "key": "pangu_cyp2d6",
+          "digest": "",
+          "tests": []
+        },
+        "pharmcat_preprocess": {
+          "key": "pharmcat_preprocess",
+          "digest": "",
+          "tests": []
+        },
+        "filter_preprocessed_vcf": {
+          "key": "filter_preprocessed_vcf",
+          "digest": "",
+          "tests": []
+        },
+        "run_pharmcat": {
+          "key": "run_pharmcat",
+          "digest": "",
+          "tests": []
+        }
+      }
+    },
+    "workflows/wdl-common/wdl/tasks/whatshap_haplotag.wdl": {
+      "key": "workflows/wdl-common/wdl/tasks/whatshap_haplotag.wdl",
+      "name": "",
+      "description": "",
+      "tasks": {
+        "whatshap_haplotag": {
+          "key": "whatshap_haplotag",
+          "digest": "",
+          "tests": []
+        }
+      }
+    },
+    "workflows/wdl-common/wdl/tasks/whatshap_phase.wdl": {
+      "key": "workflows/wdl-common/wdl/tasks/whatshap_phase.wdl",
+      "name": "",
+      "description": "",
+      "tasks": {
+        "whatshap_phase": {
+          "key": "whatshap_phase",
+          "digest": "",
+          "tests": []
+        }
+      }
+    },
+    "workflows/wdl-common/wdl/tasks/whatshap_stats.wdl": {
+      "key": "workflows/wdl-common/wdl/tasks/whatshap_stats.wdl",
+      "name": "",
+      "description": "",
+      "tasks": {
+        "whatshap_stats": {
+          "key": "whatshap_stats",
+          "digest": "",
+          "tests": []
+        }
+      }
+    },
+    "workflows/wdl-common/wdl/workflows/deepvariant/deepvariant.wdl": {
+      "key": "workflows/wdl-common/wdl/workflows/deepvariant/deepvariant.wdl",
+      "name": "",
+      "description": "",
+      "tasks": {
+        "deepvariant_make_examples": {
+          "key": "deepvariant_make_examples",
+          "digest": "",
+          "tests": []
+        },
+        "deepvariant_call_variants": {
+          "key": "deepvariant_call_variants",
+          "digest": "",
+          "tests": []
+        },
+        "deepvariant_postprocess_variants": {
+          "key": "deepvariant_postprocess_variants",
+          "digest": "",
+          "tests": []
+        }
+      }
+    },
+    "workflows/wdl-common/wdl/workflows/phase_vcf/phase_vcf.wdl": {
+      "key": "workflows/wdl-common/wdl/workflows/phase_vcf/phase_vcf.wdl",
+      "name": "",
+      "description": "",
+      "tasks": {
+        "split_vcf": {
+          "key": "split_vcf",
+          "digest": "",
+          "tests": []
+        },
+        "bcftools_concat": {
+          "key": "bcftools_concat",
+          "digest": "",
+          "tests": []
+        }
+      }
     }
   },
   "engines": {