Merge pull request #51 from kids-first/feature/mb-update-configs

🚀 Update configs
kids-first · Sep 6, 2023 · 660e352 · 660e352
2 parents bc9c9b5 + dbe5248
commit 660e352
Show file tree

Hide file tree

Showing 16 changed files with 59 additions and 58 deletions.
diff --git a/STUDY_CONFIGS/aml_sd_pet7q6f2_2018_case_meta_config.json b/STUDY_CONFIGS/aml_sd_pet7q6f2_2018_case_meta_config.json
@@ -70,13 +70,13 @@
         "dtypes": {
             "fusion": {
                 "ext": "fusions.txt",
-                "cbio_name": "data_fusions.txt",
+                "cbio_name": "data_sv.txt",
                 "meta_file_attr": {
-                    "stable_id": "fusion",
+                    "stable_id": "structural_variants",
                     "profile_name": "Predicted RNA fusions",
-                    "profile_description": "PBTA fusion data using arriba and STAR Fusion, annotated and filtered using annoFuse. Also contains DGD custom filtered fusions",
-                    "genetic_alteration_type": "FUSION",
-                    "datatype": "FUSION",
+                    "profile_description": "Fusion data using arriba and STAR Fusion, annotated and filtered using annoFuse",
+                    "genetic_alteration_type": "STRUCTURAL_VARIANT",
+                    "datatype": "SV",
                     "show_profile_in_analysis_tab": "true"
                 }
             }
@@ -105,13 +105,13 @@
     },
     "study": {
         "_comment": "see https://docs.cbioportal.org/5.1-data-loading/data-loading/file-formats#cancer-study for detailed specifics",
-        "description": "Although myeloid disorders in children may show morphologic similarities to that seen in adults, TARGET AML initiative (Meshinchi, PI) clearly demonstrated that somatic genomic and transcriptome variants are highly distinct in children and young adults, and in fact there are variants that are uniquely restricted to younger children. TARGET AML initiative, helped identify numerous somatic alterations with high therapeutic potential in younger AML patients. Clinical outcome in children with myeloid disorders have remained poor in part due to lack of deep understanding of the genomic makeup of the disease as well as the host. Comprehensive studies of the host and disease may enable more informed therapies in order to optimize targeting the leukemia while minimizing short and long term toxicities, leading to improved survival with minimal morbidities. For updates, please see here: <a href=\"https://tinyurl.com/55cxz9am\">Release Notes</a>",
+        "description": "Although myeloid disorders in children may show morphologic similarities to that seen in adults, TARGET AML initiative (Meshinchi, PI) clearly demonstrated that somatic genomic and transcriptome variants are highly distinct in children and young adults, and in fact there are variants that are uniquely restricted to younger children. TARGET AML initiative, helped identify numerous somatic alterations with high therapeutic potential in younger AML patients. Clinical outcome in children with myeloid disorders have remained poor in part due to lack of deep understanding of the genomic makeup of the disease as well as the host. Comprehensive studies of the host and disease may enable more informed therapies in order to optimize targeting the leukemia while minimizing short and long term toxicities, leading to improved survival with minimal morbidities. Additional keywords: GMKF, KF. For updates, please see here: <a href=\"https://tinyurl.com/55cxz9am\">Release Notes</a>",
         "groups": "PUBLIC",
         "cancer_study_identifier": "aml_sd_pet7q6f2_2018",
         "type_of_cancer": "aml",
         "short_name": "Kids First: Myeloid Malignancies",
         "reference_genome": "hg38",
-        "display_name": "Kids First: Germline and Somatic Variants in Myeloid Malignancies in Children (Kids First, Provisional)"
+        "display_name": "Kids First: Germline and Somatic Variants in Myeloid Malignancies in Children (phs002187, Meshinchi, Provisional)"
     },
     "cases_all": {
         "stable_id": "all",
@@ -141,14 +141,14 @@
             }
         },
         "sample_head": {
-            "table": "bix_workflows.data_clinical_sample_header"
+            "table": "template_sample_header.txt"
         },
         "sample_file": {
             "table": "brownm28_dev_schema_cbio.aml_sd_pet7q6f2_2018_data_clinical_sample",
             "out_file": "data_clinical_sample.txt"
         },
         "patient_head": {
-            "table": "bix_workflows.data_clinical_patient_header"
+            "table": "template_patient_header.txt"
         },
         "patient_file": {
             "table": "brownm28_dev_schema_cbio.aml_sd_pet7q6f2_2018_data_clinical_patient",
@@ -157,10 +157,6 @@
         "genomics_etl": {
             "table": "brownm28_dev_schema_cbio.aml_sd_pet7q6f2_2018_genomics_etl_file",
             "out_file": "cbio_file_name_id.txt"
-        },
-        "seq_center": {
-            "table": "brownm28_dev_schema_cbio.aml_sd_pet7q6f2_2018_sq_info_etl_resource",
-            "out_file": "seq_center_resource.txt"
         }
     }
 }
diff --git a/STUDY_CONFIGS/aml_sd_pet7q6f2_2018_data_processing_config.json b/STUDY_CONFIGS/aml_sd_pet7q6f2_2018_data_processing_config.json
@@ -1,9 +1,9 @@
 {
   "bedtools": "bedtools",
   "cp_only_script": "/home/ubuntu/tools/kf-cbioportal-etl/scripts/get_cbio_copy_only_num.pl",
-  "bed_genes": "/home/ubuntu/tools/kf-cbioportal-etl/REFS/GRCh38.84.gtf_genes.bed",
-  "hugo_tsv": "/home/ubuntu/tools/kf-cbioportal-etl/REFS/HUGO_EntrezID.tsv",
-  "entrez_tsv": "/home/ubuntu/tools/kf-cbioportal-etl/REFS/EntrezGeneId_HugoGeneSymbol.tsv",
+  "bed_genes": "/home/ubuntu/tools/kf-cbioportal-etl/REFS/Homo_sapiens.GRCh38.105.chr.gtf_genes.bed",
+  "hugo_tsv": "/home/ubuntu/tools/kf-cbioportal-etl/REFS/HUGO_2021-06-01_EntrezID.tsv",
+  "entrez_tsv": "/home/ubuntu/tools/kf-cbioportal-etl/REFS/EntrezGeneId_HugoGeneSymbol_2021-06-01.txt",
   "rna_ext_list": {
     "expression": "rsem.genes.results.gz",
     "fusion": "annoFuse_filter.tsv"
@@ -15,8 +15,7 @@
       "header": "/home/ubuntu/tools/kf-cbioportal-etl/REFS/maf_KF_CONSENSUS.txt"
     },
     "rsem": "RSEM_gene",
-    "fusion": "annofuse_filtered_fusions_tsv",
-    "fusion_sq_file": "seq_center_resource.txt"
+    "fusion": "annofuse_filtered_fusions_tsv"
   },
   "dl_file_type_list": ["RSEM_gene","annofuse_filtered_fusions_tsv"],
   "ens_gene_list":"/home/ubuntu/tools/kf-cbioportal-etl/REFS/gencode27_gene_list.txt",

diff --git a/STUDY_CONFIGS/aml_sd_z6mwd3h0_2018_case_meta_config.json b/STUDY_CONFIGS/aml_sd_z6mwd3h0_2018_case_meta_config.json
@@ -118,13 +118,13 @@
     },
     "study": {
         "_comment": "see https://docs.cbioportal.org/5.1-data-loading/data-loading/file-formats#cancer-study for detailed specifics",
-        "description": "Children with Down syndrome (DS), which occurs due to trisomy 21, have a 2000-fold increased risk of atrioventricular septal defects (AVSD) and a 20-fold increased risk of acute lymphoblastic leukemia (ALL), but it is not understood which genetic features of trisomy 21 are responsible for the increased risk. The objectives of this study are to determine the genetic variants underlying AVSD and ALL risk in children with DS, which builds upon our previous work suggesting having an extra copy of chromosome 21 may \"move\" the susceptibility threshold for disease in these children. Insights into the genes that drive DS-AVSD and DS-ALL may have implications for improved genetic counseling, surveillance, clinical management, and treatment strategies for these and other children who may develop AVSD or ALL. For updates, please see here: <a href=\"https://tinyurl.com/55cxz9am\">Release Notes</a>",
+        "description": "Children with Down syndrome (DS), which occurs due to trisomy 21, have a 2000-fold increased risk of atrioventricular septal defects (AVSD) and a 20-fold increased risk of acute lymphoblastic leukemia (ALL), but it is not understood which genetic features of trisomy 21 are responsible for the increased risk. The objectives of this study are to determine the genetic variants underlying AVSD and ALL risk in children with DS, which builds upon our previous work suggesting having an extra copy of chromosome 21 may \"move\" the susceptibility threshold for disease in these children. Insights into the genes that drive DS-AVSD and DS-ALL may have implications for improved genetic counseling, surveillance, clinical management, and treatment strategies for these and other children who may develop AVSD or ALL. Additional keywords: GMKF, KF. For updates, please see here: <a href=\"https://tinyurl.com/55cxz9am\">Release Notes</a>",
         "groups": "PUBLIC",
         "cancer_study_identifier": "aml_sd_z6mwd3h0_2018",
         "type_of_cancer": "aml",
         "short_name": "Genomic Analysis of CHD and ALL in Children with Down Syndrome",
         "reference_genome": "hg38",
-        "display_name": "Genomic Analysis of Congenital Heart Defects and Acute Lymphoblastic Leukemia in Children with Down Syndrome (Kids First, Provisional)"
+        "display_name": "Kids First and INCLUDE: Down Syndrome, Heart Defects, and Acute Lymphoblastic Leukemia (phs002330, Lupo, Provisional)"
     },
     "cases_3way_complete": {
         "stable_id": "3way_complete",
@@ -178,14 +178,14 @@
             }
         },
         "sample_head": {
-            "table": "bix_workflows.data_clinical_sample_header"
+            "table": "template_sample_header.txt"
         },
         "sample_file": {
             "table": "brownm28_dev_schema_cbio.aml_sd_z6mwd3h0_2018_data_clinical_sample",
             "out_file": "data_clinical_sample.txt"
         },
         "patient_head": {
-            "table": "bix_workflows.data_clinical_patient_header"
+            "table": "template_patient_header.txt"
         },
         "patient_file": {
             "table": "brownm28_dev_schema_cbio.aml_sd_z6mwd3h0_2018_data_clinical_patient",

diff --git a/STUDY_CONFIGS/aml_sd_z6mwd3h0_2018_data_processing_config.json b/STUDY_CONFIGS/aml_sd_z6mwd3h0_2018_data_processing_config.json
@@ -1,9 +1,9 @@
 {
   "bedtools": "bedtools",
   "cp_only_script": "/home/ubuntu/tools/kf-cbioportal-etl/scripts/get_cbio_copy_only_num.pl",
-  "bed_genes": "/home/ubuntu/tools/kf-cbioportal-etl/REFS/GRCh38.84.gtf_genes.bed",
-  "hugo_tsv": "/home/ubuntu/tools/kf-cbioportal-etl/REFS/HUGO_EntrezID.tsv",
-  "entrez_tsv": "/home/ubuntu/tools/kf-cbioportal-etl/REFS/EntrezGeneId_HugoGeneSymbol.tsv",
+  "bed_genes": "/home/ubuntu/tools/kf-cbioportal-etl/REFS/Homo_sapiens.GRCh38.105.chr.gtf_genes.bed",
+  "hugo_tsv": "/home/ubuntu/tools/kf-cbioportal-etl/REFS/HUGO_2021-06-01_EntrezID.tsv",
+  "entrez_tsv": "/home/ubuntu/tools/kf-cbioportal-etl/REFS/EntrezGeneId_HugoGeneSymbol_2021-06-01.txt",
   "rna_ext_list": {
     "expression": "rsem.genes.results.gz",
     "fusion": "annoFuse_filter.tsv"
@@ -17,7 +17,7 @@
     "_comment": "edit the values based on existing/anticipated source file locations, relative to working directory of the script being run",
     "mafs": {
       "kf": "annotated_public_outputs",
-      "header": "/home/ubuntu/tools/kf-cbioportal-etl/REFS/maf_KF_CONSENSUS.txt"
+      "header": "/home/ubuntu/tools/kf-cbioportal-etl/REFS/maf_KF_CONSENSUS_r105.txt"
     },
     "cnvs": {
       "pval": "ctrlfreec_pval",

diff --git a/STUDY_CONFIGS/chdm_phs001643_2018_case_meta_config.json b/STUDY_CONFIGS/chdm_phs001643_2018_case_meta_config.json
@@ -226,14 +226,14 @@
             }
         },
         "sample_head": {
-            "table": "bix_workflows.data_clinical_sample_header"
+            "table": "template_sample_header.txt"
         },
         "sample_file": {
             "table": "prod_cbio.chdm_phs001643_2018_data_clinical_sample",
             "out_file": "data_clinical_sample.txt"
         },
         "patient_head": {
-            "table": "bix_workflows.data_clinical_patient_header"
+            "table": "template_patient_header.txt"
         },
         "patient_file": {
             "table": "prod_cbio.chdm_phs001643_2018_data_clinical_patient",

diff --git a/STUDY_CONFIGS/chdm_sd_7spqtt8m_case_meta_config.json b/STUDY_CONFIGS/chdm_sd_7spqtt8m_case_meta_config.json
@@ -231,14 +231,14 @@
             }
         },
         "sample_head": {
-            "table": "bix_workflows.data_clinical_sample_header"
+            "table": "template_sample_header.txt"
         },
         "sample_file": {
             "table": "prod_cbio.chdm_sd_7spqtt8m_data_clinical_sample",
             "out_file": "data_clinical_sample.txt"
         },
         "patient_head": {
-            "table": "bix_workflows.data_clinical_patient_header"
+            "table": "template_patient_header.txt"
         },
         "patient_file": {
             "table": "prod_cbio.chdm_sd_7spqtt8m_data_clinical_patient",

diff --git a/STUDY_CONFIGS/meningioma_rna_publication_meta_config.json b/STUDY_CONFIGS/meningioma_rna_publication_meta_config.json
@@ -123,14 +123,14 @@
             }
         },
         "sample_head": {
-            "table": "bix_workflows.data_clinical_sample_header"
+            "table": "template_sample_header.txt"
         },
         "sample_file": {
             "table": "brownm28_dev_schema.delme_lung_mito_rna_data_clinical_sample",
             "out_file": "data_clinical_sample.txt"
         },
         "patient_head": {
-            "table": "bix_workflows.data_clinical_patient_header"
+            "table": "template_patient_header.txt"
         },
         "patient_file": {
             "table": "brownm28_dev_schema.delme_lung_mito_rna_data_clinical_patient",

diff --git a/STUDY_CONFIGS/oligo_nation_case_meta_config.json b/STUDY_CONFIGS/oligo_nation_case_meta_config.json
@@ -183,14 +183,14 @@
             }
         },
         "sample_head": {
-            "table": "bix_workflows.data_clinical_sample_header"
+            "table": "template_sample_header.txt"
         },
         "sample_file": {
             "table": "prod_cbio.oligo_nation_data_clinical_sample",
             "out_file": "data_clinical_sample.txt"
         },
         "patient_head": {
-            "table": "bix_workflows.data_clinical_patient_header"
+            "table": "template_patient_header.txt"
         },
         "patient_file": {
             "table": "prod_cbio.oligo_nation_data_clinical_patient",

diff --git a/STUDY_CONFIGS/open_chordoma_case_meta_config.json b/STUDY_CONFIGS/open_chordoma_case_meta_config.json
@@ -235,14 +235,14 @@
             }
         },
         "sample_head": {
-            "table": "bix_workflows.data_clinical_sample_header"
+            "table": "template_sample_header.txt"
         },
         "sample_file": {
             "table": "brownm28_dev_schema_cbio.open_chordoma_data_clinical_sample",
             "out_file": "data_clinical_sample.txt"
         },
         "patient_head": {
-            "table": "bix_workflows.data_clinical_patient_header"
+            "table": "template_patient_header.txt"
         },
         "patient_file": {
             "table": "brownm28_dev_schema_cbio.open_chordoma_data_clinical_patient",

diff --git a/STUDY_CONFIGS/os_sd_zxjffmef_2015_case_meta_config.json b/STUDY_CONFIGS/os_sd_zxjffmef_2015_case_meta_config.json
@@ -118,13 +118,13 @@
     },
     "study": {
         "_comment": "see https://docs.cbioportal.org/5.1-data-loading/data-loading/file-formats#cancer-study for detailed specifics",
-        "description": "Although the survival of children with relapsed osteosarcoma is very poor, little is known about the etiology of treatment failure in this disease. The purpose of this project is to perform whole genome sequencing on serial samples from patients with osteosarcoma obtained before treatment, after treatment, and at relapse in order to identify the mutations and pathways that are drivers of drug resistance. If successful, our results may help identify patients at high risk for treatment failure and may yield new treatments for children who cannot currently be cured. For updates, please see here: <a href=\"https://tinyurl.com/55cxz9am\">Release Notes</a>",
+        "description": "Although the survival of children with relapsed osteosarcoma is very poor, little is known about the etiology of treatment failure in this disease. The purpose of this project is to perform whole genome sequencing on serial samples from patients with osteosarcoma obtained before treatment, after treatment, and at relapse in order to identify the mutations and pathways that are drivers of drug resistance. If successful, our results may help identify patients at high risk for treatment failure and may yield new treatments for children who cannot currently be cured. Additional keywords: GMKF, KF. For updates, please see here: <a href=\"https://tinyurl.com/55cxz9am\">Release Notes</a>",
         "groups": "PUBLIC",
         "cancer_study_identifier": "os_sd_zxjffmef_2015",
         "type_of_cancer": "os",
         "short_name": "An Integrated Analysis of Treatment Failure in Pediatric OS",
         "reference_genome": "hg38",
-        "display_name": "An Integrated Clinical and Genomic Analysis of Treatment Failure in Pediatric Osteosarcoma (Kids First, Provisional)"
+        "display_name": "Kids First: An Integrated Clinical and Genomic Analysis of Treatment Failure in Pediatric Osteosarcoma (phs001714, Onel, Provisional)"
     },
     "cases_3way_complete": {
         "stable_id": "3way_complete",
@@ -178,14 +178,14 @@
             }
         },
         "sample_head": {
-            "table": "bix_workflows.data_clinical_sample_header"
+            "table": "template_sample_header.txt"
         },
         "sample_file": {
             "table": "brownm28_dev_schema_cbio.os_sd_zxjffmef_2015_data_clinical_sample",
             "out_file": "data_clinical_sample.txt"
         },
         "patient_head": {
-            "table": "bix_workflows.data_clinical_patient_header"
+            "table": "template_patient_header.txt"
         },
         "patient_file": {
             "table": "brownm28_dev_schema_cbio.os_sd_zxjffmef_2015_data_clinical_patient",

diff --git a/STUDY_CONFIGS/os_sd_zxjffmef_2015_data_processing_config.json b/STUDY_CONFIGS/os_sd_zxjffmef_2015_data_processing_config.json
@@ -1,30 +1,25 @@
 {
   "bedtools": "bedtools",
   "cp_only_script": "/home/ubuntu/tools/kf-cbioportal-etl/scripts/get_cbio_copy_only_num.pl",
-  "bed_genes": "/home/ubuntu/tools/kf-cbioportal-etl/REFS/GRCh38.84.gtf_genes.bed",
-  "hugo_tsv": "/home/ubuntu/tools/kf-cbioportal-etl/REFS/HUGO_EntrezID.tsv",
-  "entrez_tsv": "/home/ubuntu/tools/kf-cbioportal-etl/REFS/EntrezGeneId_HugoGeneSymbol.tsv",
-  "rna_ext_list": {
-    "expression": "rsem.genes.results.gz",
-    "fusion": "annoFuse_filter.tsv"
-  },
+  "bed_genes": "/home/ubuntu/tools/kf-cbioportal-etl/REFS/Homo_sapiens.GRCh38.105.chr.gtf_genes.bed",
+  "hugo_tsv": "/home/ubuntu/tools/kf-cbioportal-etl/REFS/HUGO_2021-06-01_EntrezID.tsv",
+  "entrez_tsv": "/home/ubuntu/tools/kf-cbioportal-etl/REFS/EntrezGeneId_HugoGeneSymbol_2021-06-01.txt",
   "dna_ext_list": {
-    "mutation": "consensus_somatic.vep.maf", 
+    "mutation": "consensus_somatic.norm.annot.public.maf", 
     "copy_number": "controlfreec.CNVs.p.value.txt",
     "seg": "controlfreec.seg"
   },
   "file_loc_defs": {
     "_comment": "edit the values based on existing/anticipated source file locations, relative to working directory of the script being run",
     "mafs": {
       "kf": "annotated_public_outputs",
-      "header": "/home/ubuntu/tools/kf-cbioportal-etl/REFS/maf_KF_CONSENSUS.txt"
+      "header": "/home/ubuntu/tools/kf-cbioportal-etl/REFS/maf_KF_CONSENSUS_r105.txt"
     },
     "cnvs": {
       "pval": "ctrlfreec_pval",
       "info": "ctrlfreec_info",
       "seg": "ctrlfreec_bam_seg"
     }
-
   },
   "dl_file_type_list": ["annotated_public_outputs",
     "ctrlfreec_pval","ctrlfreec_info","ctrlfreec_bam_seg"],

diff --git a/STUDY_CONFIGS/pbta_all_case_meta_config.json b/STUDY_CONFIGS/pbta_all_case_meta_config.json
@@ -238,7 +238,18 @@
                 "table": "bix_genomics_file.sd_8y99qzjj-genomics_file_manifest",
                 "file_type": ["RSEM_gene","annofuse_filtered_fusions_tsv","annotated_public_outputs","ctrlfreec_pval","ctrlfreec_info","ctrlfreec_bam_seg"],
                 "out_file":  "pnoc_genomics_file_manifest.txt"
+            },
+            "oligo": {
+                "table": "bix_genomics_file.sd_bhjxbdqk_oligo-genomics_file_manifest",
+                "file_type": ["RSEM_gene","annofuse_filtered_fusions_tsv","annotated_public_outputs","ctrlfreec_pval","ctrlfreec_info","ctrlfreec_bam_seg"],
+                "out_file": "oligo_nation_genomics_file_manifest.txt"
+            },
+            "mioncoseq": {
+                "table": "bix_genomics_file.sd_bhjxbdqk_mioncoseq-genomics_file_manifest",
+                "file_type": ["RSEM_gene","annofuse_filtered_fusions_tsv","annotated_public_outputs","ctrlfreec_pval","ctrlfreec_info","ctrlfreec_bam_seg"],
+                "out_file": "mioncoseq_genomics_file_manifest.txt"
             }
+
         },
         "sample_head": {
             "table": "template_sample_header.txt"

diff --git a/STUDY_CONFIGS/pbta_all_treatment_meta_config.json b/STUDY_CONFIGS/pbta_all_treatment_meta_config.json
@@ -245,14 +245,14 @@
             }
         },
         "sample_head": {
-            "table": "bix_workflows.data_clinical_sample_header"
+            "table": "template_sample_header.txt"
         },
         "sample_file": {
             "table": "brownm28_dev_schema_cbio.pbta_all_data_clinical_sample",
             "out_file": "data_clinical_sample.txt"
         },
         "patient_head": {
-            "table": "bix_workflows.data_clinical_patient_header"
+            "table": "template_patient_header.txt"
         },
         "patient_file": {
             "table": "brownm28_dev_schema_cbio.pbta_all_data_clinical_patient",