nci-cidc-schemas 0.27.20__tar.gz → 0.27.22__tar.gz

{nci_cidc_schemas-0.27.20/nci_cidc_schemas.egg-info → nci_cidc_schemas-0.27.22}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: nci_cidc_schemas
-Version: 0.27.20
+Version: 0.27.22
 Summary: The CIDC data model and tools for working with it.
 Home-page: https://github.com/NCI-CIDC/cidc-schemas
 Author: NCI

{nci_cidc_schemas-0.27.20 → nci_cidc_schemas-0.27.22}/cidc_schemas/__init__.py

@@ -2,4 +2,4 @@
 
 __author__ = """NCI"""
 __email__ = "nci-cidc-tools-admin@mail.nih.gov"
-__version__ = "0.27.20"
+__version__ = "0.27.22"

{nci_cidc_schemas-0.27.20 → nci_cidc_schemas-0.27.22}/cidc_schemas/prism/constants.py

@@ -26,6 +26,7 @@ SUPPORTED_ASSAYS = [
     "ctdna",
     "microbiome",
     "mihc",
+    "scrnaseq",
 ]
 
 SUPPORTED_SHIPPING_MANIFESTS = [
@@ -93,6 +94,7 @@ ASSAY_TO_FILEPATH: Dict[str, str] = {
             "nanostring",
             "olink",
             "mihc",
+            "scrnaseq",
         ]
     },
 }

{nci_cidc_schemas-0.27.20 → nci_cidc_schemas-0.27.22}/cidc_schemas/schemas/assays/components/available_assays.json

@@ -130,6 +130,17 @@
         "idRef": "batch_id"
       }
     },
+    "scrnaseq": {
+      "type": "array",
+      "description": "scRNASeq",
+      "items": {
+        "$ref": "assays/scrnaseq_assay.json"
+      },
+      "mergeStrategy": "arrayMergeById",
+      "mergeOptions": {
+        "idRef": "batch_id"
+      }
+    },
     "nanostring": {
       "type": "array",
       "description": "Nanostring",

nci_cidc_schemas-0.27.22/cidc_schemas/schemas/assays/components/scrnaseq_input.json

@@ -0,0 +1,39 @@
+{
+  "$schema": "metaschema/strict_meta_schema.json#",
+  "$id": "scrnaseq_input",
+  "title": "scRNAseq Input Files and Data",
+  "type": "object",
+  "description": "scRNAseq assay input files and data",
+  "inheritableBase": true,
+  "properties": {
+    "lane": {
+      "description": "Lane used for this run",
+      "type": "string"
+    },
+    "read_1_file": {
+      "description": "Contains the cell barcode and unique molecular identifier",
+      "$ref": "artifacts/artifact_gz.json"
+    },
+    "read_2_file": {
+      "description": "Contains the cDNA (transcript) sequence",
+      "$ref": "artifacts/artifact_gz.json"
+    },
+    "library_yield_ng": {
+      "description": "Resulting yield (in ng) from library construction.",
+      "type": "number"
+    },
+    "rin": {
+      "description": "RNA integrity number. (0-10)",
+      "type": "number",
+      "minimum": 0,
+      "maximum": 10
+    }
+  },
+  "required": [
+    "lane",
+    "read_1_file",
+    "read_2_file",
+    "library_yield_ng",
+    "rin"
+  ]
+}

nci_cidc_schemas-0.27.22/cidc_schemas/schemas/assays/components/scrnaseq_preamble.json

@@ -0,0 +1,65 @@
+{
+  "$schema": "metaschema/strict_meta_schema.json#",
+  "$id": "scrnaseq_preamble",
+  "title": "ScRNA Preamble",
+  "description": "Preamble information in an scRNAseq assay",
+  "type": "object",
+  "inheritableBase": true,
+  "properties": {
+    "azimuth_reference": {
+      "description": "Cell type annotation will be performed with this Azimuth reference for all samples in the batch.",
+      "type": "string",
+      "enum": [
+        "adiposeref",
+        "bonemarrowref",
+        "fetusref",
+        "heartref",
+        "humancortexref",
+        "kidneyref",
+        "lungref",
+        "pancreasref",
+        "pbmcref",
+        "tonsilref",
+        "none"
+      ]
+    },
+    "tabula_sapiens_reference": {
+      "description": "Cell type annotation will be performed with this Tabula Sapiens reference for all samples in the batch.",
+      "type": "string",
+      "enum": [
+        "bladder",
+        "blood",
+        "bone marrow",
+        "eye",
+        "fat",
+        "immune",
+        "mammary",
+        "prostrate",
+        "skin",
+        "endothelial",
+        "none"
+      ]
+    },
+    "assay_chemistry_version": {
+      "description": "The specific version of the 10x Genomics single-cell gene expression assay chemistry used to prepare and sequence a sample.",
+      "type": "string",
+      "enum": [
+        "Single Cell 3' v2",
+        "Single Cell 3' v3",
+        "Single Cell 3' v3.1",
+        "Single Cell 3' HT v3.1",
+        "Single Cell 3' v4"
+      ]
+    },
+    "barcode_include_file_version": {
+      "description": "Barcode inclusion file version for the 10x Genomics single-cell gene expression assay. Refer to https://kb.10xgenomics.com/hc/en-us/articles/115004506263-What-is-a-barcode-whitelist for more information. '2' for 737K-august-2016.txt (Single Cell 3' v2), '3' for 3M-february-2018.txt.gz (Single Cell 3' v3, Single Cell 3' v3.1, Single Cell 3' HT v3.1), '4' for 3M-3pgex-may-2023.txt.gz (Single Cell 3' v4)",
+      "type": "string",
+      "enum": ["2","3","4"]
+    },
+
+    "total_barcode_and_umi_length_in_read": {
+      "description": "Should be a a positive integer around 26-28 bp. Equals to the sum of the cell barcode and the unique molecular identifier (UMI) in read 1.",
+      "type": "integer"
+    }
+  }
+}

nci_cidc_schemas-0.27.22/cidc_schemas/schemas/assays/scrnaseq_assay.json

@@ -0,0 +1,74 @@
+{
+    "$schema": "metaschema/strict_meta_schema.json#",
+    "$id": "scrnaseq_assay",
+    "title": "scRNAseq Assay",
+    "type": "object",
+    "$comment": "description used in docs/index.html and docs/assays.scrnaseq.html",
+    "description": "Single Cell RNA Sequencing (scRNAseq) assay",
+    "additionalProperties": false,
+    "definitions": {
+        "record": {
+            "type": "object",
+            "mergeStrategy": "objectMerge",
+            "description": "A single data record from a scRNAseq assay.",
+            "additionalProperties": false,
+            "required": [
+                "cimac_id",
+                "details"
+            ],
+            "properties": {
+                "cimac_id": {
+                    "description": "Id of an sample within this clinical trial, that this assay record is based upon.",
+                    "$ref": "sample.json#properties/cimac_id"
+                },
+                "details": {
+                    "$ref": "assays/components/scrnaseq_input.json"
+                }
+            }
+        }
+    },
+    "properties": {
+        "assay_creator": {
+            "$ref": "assays/components/assay_core.json#properties/assay_creator"
+        },
+        "batch_id": {
+            "description": "Batch identification number. Unique to every upload.",
+            "type": "string"
+        },
+        "azimuth_reference": {
+            "$ref": "assays/components/scrnaseq_preamble.json#properties/azimuth_reference"
+        },
+        "tabula_sapiens_reference": {
+            "$ref": "assays/components/scrnaseq_preamble.json#properties/tabula_sapiens_reference"
+        },
+        "assay_chemistry_version": {
+            "$ref": "assays/components/scrnaseq_preamble.json#properties/assay_chemistry_version"
+        },
+        "barcode_include_file_version": {
+            "$ref": "assays/components/scrnaseq_preamble.json#properties/barcode_include_file_version"
+        },
+        "total_barcode_and_umi_length_in_read": {
+            "$ref": "assays/components/scrnaseq_preamble.json#properties/total_barcode_and_umi_length_in_read"
+        },
+
+        "samples_metadata_csv": {"$ref": "artifacts/artifact_csv.json"},
+
+
+        "records": {
+            "type": "array",
+            "description": "A single data record from a scrnaseq assay.",
+            "items": {
+                "$ref": "assays/scrnaseq_assay.json#definitions/record"
+            },
+            "mergeStrategy": "append"
+        }
+    },
+    "required": [
+        "assay_creator",
+        "batch_id",
+        "assay_chemistry_version",
+        "barcode_include_file_version",
+        "samples_metadata_csv",
+        "records"
+    ]
+}

{nci_cidc_schemas-0.27.20 → nci_cidc_schemas-0.27.22}/cidc_schemas/schemas/sample.json

@@ -6,7 +6,7 @@
 
 	"$comment": "description used in docs/index.html and docs/sample.html",
     "description": "Any material sample taken from a biological entity for testing, diagnostic, propagation, treatment or research purposes, including a sample obtained from a living organism or taken from the biological object after halting of all its life functions. Biospecimen can contain one or more components including but not limited to cellular molecules, cells, tissues, organs, body fluids, embryos, and body excretory products.",
-    
+
     "properties": {
         "cimac_id": {
             "description": "Specimen identifier assigned by the CIMAC-CIDC Network. Formatted as CTTTPPPSS.AA for trial code TTT, participant PPP, sample SS, and aliquot AA.",

{nci_cidc_schemas-0.27.20 → nci_cidc_schemas-0.27.22}/cidc_schemas/schemas/shipping_core.json

@@ -60,7 +60,7 @@
         "Nanostring",
         "NULISA",
         "MIBI",
-        "scRNA-Seq",
+        "scRNAseq",
         "snRNA-Seq",
         "Visium"
       ]

nci_cidc_schemas-0.27.22/cidc_schemas/schemas/templates/assays/scrnaseq_template.json

@@ -0,0 +1,97 @@
+{    "title": "scRNAseq assay template",
+    "description": "Single Cell RNA Sequencing (scRNAseq) submission.",
+    "prism_template_root_object_schema": "assays/scrnaseq_assay.json",
+    "prism_template_root_object_pointer": "/assays/scrnaseq/0",
+    "properties":
+    {
+        "worksheets":
+        {
+            "scRNAseq":
+            {
+                "preamble_rows":
+                {
+                    "protocol identifier":
+                    {
+                        "merge_pointer": "3/protocol_identifier",
+                        "type_ref": "clinical_trial.json#properties/protocol_identifier"
+                    },
+                    "assay creator":
+                    {
+                        "merge_pointer": "0/assay_creator",
+                        "type_ref": "assays/scrnaseq_assay.json#properties/assay_creator"
+                    },
+                    "batch id": {
+                        "merge_pointer": "0/batch_id",
+                        "type_ref": "assays/scrnaseq_assay.json#properties/batch_id"
+                    },
+                    "azimuth reference": {
+                        "merge_pointer": "0/azimuth_reference",
+                        "type_ref": "assays/components/scrnaseq_preamble.json#properties/azimuth_reference",
+                        "allow_empty": true
+                    },
+                    "tabula sapiens reference": {
+                        "merge_pointer": "0/tabula_sapiens_reference",
+                        "type_ref": "assays/components/scrnaseq_preamble.json#properties/tabula_sapiens_reference",
+                        "allow_empty": true
+                    },
+                    "assay chemistry version": {
+                        "merge_pointer": "0/assay_chemistry_version",
+                        "type_ref": "assays/components/scrnaseq_preamble.json#properties/assay_chemistry_version"
+                    },
+                    "barcode include file version": {
+                        "merge_pointer": "0/barcode_include_file_version",
+                        "type_ref": "assays/components/scrnaseq_preamble.json#properties/barcode_include_file_version"
+                    },
+                    "total barcode and umi length in read 1": {
+                        "merge_pointer": "0/total_barcode_and_umi_length_in_read",
+                        "type_ref": "assays/components/scrnaseq_preamble.json#properties/total_barcode_and_umi_length_in_read"
+                    },
+                    "sample metadata csv": {
+                        "merge_pointer": "0/samples_metadata_csv",
+                        "gcs_uri_format": "{protocol identifier}/scrnaseq/{batch id}/samples_metadata.csv",
+                        "is_artifact": 1,
+                        "type_ref": "assays/components/local_file.json#properties/file_path"
+                    }
+                },
+                "prism_data_object_pointer": "/records/-",
+                "data_columns":
+                {
+                    "Samples":
+                    {
+                        "Cimac ID":
+                        {
+                            "merge_pointer": "0/cimac_id",
+                            "type_ref": "sample.json#properties/cimac_id"
+                        },
+                        "Lane": {
+                            "merge_pointer": "0/details/lane",
+                            "type": "string"
+                        },
+                        "Read 1": {
+                            "merge_pointer": "0/details/read_1_file",
+                            "is_artifact": 1,
+                            "gcs_uri_format": "{protocol identifier}/scrnaseq/{batch id}/{cimac id}/read_1.gz",
+                            "type_ref": "assays/components/local_file.json#properties/file_path"
+                        },
+                        "Read 2": {
+                            "merge_pointer": "0/details/read_2_file",
+                            "is_artifact": 1,
+                            "gcs_uri_format": "{protocol identifier}/scrnaseq/{batch id}/{cimac id}/read_1.gz",
+                            "type_ref": "assays/components/local_file.json#properties/file_path"
+                        },
+                        "library yield ng": {
+                            "merge_pointer": "0/details/library_yield_ng",
+                            "type_ref": "assays/components/scrnaseq_input.json#properties/library_yield_ng",
+                            "allow_empty": true
+                        },
+                        "rin": {
+                            "merge_pointer": "0/details/rin",
+                            "type_ref": "assays/components/scrnaseq_input.json#properties/rin",
+                            "allow_empty": true
+                        }
+                    }
+                }
+            }
+        }
+    }
+}

{nci_cidc_schemas-0.27.20 → nci_cidc_schemas-0.27.22}/cidc_schemas/schemas/templates/manifests/normal_blood_dna_template.json

@@ -219,7 +219,8 @@
                         },
                         "type of tumor sample": {
                             "merge_pointer": "/type_of_tumor_sample",
-                            "type_ref": "sample.json#properties/type_of_tumor_sample"
+                            "type_ref": "sample.json#properties/type_of_tumor_sample",
+                            "allow_empty": true
                         },
                         "sample collection procedure": {
                             "merge_pointer": "/sample_collection_procedure",

{nci_cidc_schemas-0.27.20 → nci_cidc_schemas-0.27.22/nci_cidc_schemas.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: nci_cidc_schemas
-Version: 0.27.20
+Version: 0.27.22
 Summary: The CIDC data model and tools for working with it.
 Home-page: https://github.com/NCI-CIDC/cidc-schemas
 Author: NCI

{nci_cidc_schemas-0.27.20 → nci_cidc_schemas-0.27.22}/nci_cidc_schemas.egg-info/SOURCES.txt

@@ -83,6 +83,7 @@ cidc_schemas/schemas/assays/nanostring_assay.json
 cidc_schemas/schemas/assays/olink_assay.json
 cidc_schemas/schemas/assays/rna_assay-v0.json
 cidc_schemas/schemas/assays/rna_assay.json
+cidc_schemas/schemas/assays/scrnaseq_assay.json
 cidc_schemas/schemas/assays/tcr_analysis.json
 cidc_schemas/schemas/assays/tcr_assay.json
 cidc_schemas/schemas/assays/wes_analysis.json
@@ -111,6 +112,8 @@ cidc_schemas/schemas/assays/components/ngs_assay_core.json
 cidc_schemas/schemas/assays/components/ngs_assay_record.json
 cidc_schemas/schemas/assays/components/reads_core.json
 cidc_schemas/schemas/assays/components/reads_with_index.json
+cidc_schemas/schemas/assays/components/scrnaseq_input.json
+cidc_schemas/schemas/assays/components/scrnaseq_preamble.json
 cidc_schemas/schemas/assays/components/imaging/mif_entry.json
 cidc_schemas/schemas/assays/components/imaging/mif_export.json
 cidc_schemas/schemas/assays/components/imaging/mif_input.json
@@ -150,6 +153,7 @@ cidc_schemas/schemas/templates/assays/nanostring_template.json
 cidc_schemas/schemas/templates/assays/olink_template.json
 cidc_schemas/schemas/templates/assays/rna_bam_template.json
 cidc_schemas/schemas/templates/assays/rna_fastq_template.json
+cidc_schemas/schemas/templates/assays/scrnaseq_template.json
 cidc_schemas/schemas/templates/assays/tcr_adaptive_template.json
 cidc_schemas/schemas/templates/assays/tcr_fastq_template.json
 cidc_schemas/schemas/templates/assays/wes_bam_template.json

{nci_cidc_schemas-0.27.20 → nci_cidc_schemas-0.27.22}/tests/prism/cidc_test_data/assay_data.py

@@ -3683,6 +3683,116 @@ def nanostring() -> PrismTestData:
     )
 
 
+@assay_data_generator
+def scrnaseq() -> PrismTestData:
+    upload_type = "scrnaseq"
+    prismify_args = get_prismify_args(upload_type)
+    prismify_patch = {
+        "protocol_identifier": "test_prism_trial_id",
+        "assays": {
+            "scrnaseq": [
+                {
+                    "assay_creator": "DFCI",
+                    "batch_id": "test_batch",
+                    "azimuth_reference": "adiposeref",
+                    "tabula_sapiens_reference": "bladder",
+                    "assay_chemistry_version": "Single Cell 3' v2",
+                    "barcode_include_file_version": "2",
+                    "total_barcode_and_umi_length_in_read": 13456789012,
+                    "samples_metadata_csv": {
+                        "upload_placeholder": "d658b480-ed78-4717-b622-3e222333"
+                    },
+                    "records": [
+                        {
+                            "cimac_id": "CTTTPP121.01",
+                            "details": {
+                                "lane": "1",
+                                "read_1_file": {
+                                    "upload_placeholder": "4e9d0a47-90dc-4134-9ad6-3e3dd836111"
+                                },
+                                "read_2_file": {
+                                    "upload_placeholder": "4e9d0a47-90dc-4134-9ad6-3e3dd836222"
+                                },
+                                "library_yield_ng": 32,
+                                "rin": 3,
+                            },
+                        },
+                        {
+                            "cimac_id": "CTTTPP122.01",
+                            "details": {
+                                "lane": "2",
+                                "read_1_file": {
+                                    "upload_placeholder": "4e9d0a47-90dc-4134-9ad6-3e3dd836333"
+                                },
+                                "read_2_file": {
+                                    "upload_placeholder": "4e9d0a47-90dc-4134-9ad6-3e3dd836444"
+                                },
+                                "library_yield_ng": 33,
+                                "rin": 8,
+                            },
+                        },
+                    ],
+                }
+            ]
+        },
+    }
+    upload_entries = [
+        LocalFileUploadEntry(
+            local_path="samples.csv",
+            upload_placeholder="d658b480-ed78-4717-b622-3e222333",
+            gs_key="test_prism_trial_id/scrnaseq/test_batch/samples_metadata.csv",
+            metadata_availability=False,
+            allow_empty=False,
+        ),
+        LocalFileUploadEntry(
+            local_path="1_1.gz",
+            upload_placeholder="4e9d0a47-90dc-4134-9ad6-3e3dd836111",
+            gs_key="test_prism_trial_id/scrnaseq/test_batch/CTTTPP121.01/read_1.gz",
+            metadata_availability=False,
+            allow_empty=False,
+        ),
+        LocalFileUploadEntry(
+            local_path="1_2.gz",
+            upload_placeholder="4e9d0a47-90dc-4134-9ad6-3e3dd836222",
+            gs_key="test_prism_trial_id/scrnaseq/test_batch/CTTTPP121.01/read_2.gz",
+            metadata_availability=False,
+            allow_empty=False,
+        ),
+        LocalFileUploadEntry(
+            local_path="2_1.gz",
+            upload_placeholder="4e9d0a47-90dc-4134-9ad6-3e3dd836333",
+            gs_key="test_prism_trial_id/scrnaseq/test_batch/CTTTPP122.01/read_1.gz",
+            metadata_availability=False,
+            allow_empty=False,
+        ),
+        LocalFileUploadEntry(
+            local_path="2_2.gz",
+            upload_placeholder="4e9d0a47-90dc-4134-9ad6-3e3dd836444",
+            gs_key="test_prism_trial_id/scrnaseq/test_batch/CTTTPP122.01/read_2.gz",
+            metadata_availability=False,
+            allow_empty=False,
+        ),
+    ]
+
+    cimac_ids = [
+        record["cimac_id"]
+        for batch in prismify_patch["assays"]["scrnaseq"]
+        for record in batch["records"]
+    ]
+    base_trial = get_test_trial(cimac_ids)
+
+    target_trial = copy_dict_with_branch(base_trial, prismify_patch, "assays")
+
+    return PrismTestData(
+        upload_type,
+        prismify_args,
+        prismify_patch,
+        upload_entries,
+        base_trial,
+        target_trial,
+    )
+
+
 @assay_data_generator
 def ctdna() -> PrismTestData:
     upload_type = "ctdna"