National Center for Biotechnology Information (NCBI) Prokaryote API

openapi: 3.0.1
info:
  title: NCBI Datasets BioSample Prokaryote API
  version: v2
  description: '### NCBI Datasets is a resource that lets you easily gather data from NCBI.

    The NCBI Datasets version 2 API is updated often to add new features, fix bugs, and enhance usability.

    '
servers:
- url: https://api.ncbi.nlm.nih.gov/datasets/v2
security:
- ApiKeyAuthHeader: []
tags:
- name: Prokaryote
  description: '#### Options to download prokaryote gene data, including the associated sequence and metadata.

    These gene services allow you to download gene and protein sequence, and metadata as a prokaryote gene data package, for prokaryotic [RefSeq non-redundant proteins](https://www.ncbi.nlm.nih.gov/refseq/about/nonredundantproteins/).

    '
paths:
  /protein/accession/{accessions}/download:
    get:
      summary: Get a prokaryote gene data package by RefSeq protein accession
      description: Download a prokaryote gene data package including sequence, annotation and data reports by RefSeq non-redundant protein accession.
      tags:
      - Prokaryote
      operationId: download_prokaryote_gene_package
      responses:
        default:
          description: An unexpected error response.
          content:
            text/plain:
              schema:
                $ref: '#/components/schemas/rpcStatus'
        '200':
          description: A successful response
          content:
            application/zip:
              schema:
                format: binary
                type: string
                description: Zip compressed stream
      parameters:
      - name: accessions
        description: One or more RefSeq non-redundant protein accessions (preceded by `WP_`)
        in: path
        required: true
        schema:
          type: array
          items:
            type: string
        examples:
          example-0:
            value: WP_015878339.1
            summary: DNA-deoxyinosine glycosylase
          example-1:
            value: WP_001040085.1
            summary: type II CRISPR RNA-guided endonuclease Cas9
      - name: include_annotation_type
        description: Specify which sequence files to include in the data package.
        in: query
        required: false
        schema:
          type: array
          items:
            $ref: '#/components/schemas/v2Fasta'
        examples:
          example-0:
            value: FASTA_GENE
            summary: Include gene sequence file, which includes gene sequences for all genomes on which the protein is annotated
          example-1:
            value:
            - FASTA_GENE
            - FASTA_PROTEIN
            summary: Include gene and protein sequence files
      - name: gene_flank_config.length
        in: query
        required: false
        schema:
          type: integer
      - name: taxon
        description: Limit gene sequences and the genome assemblies described by the [annotation report file](https://www.ncbi.nlm.nih.gov/datasets/docs/v2/reference-docs/data-reports/prokaryote-gene-location/) to the specified NCBI Taxonomy ID or name (common or scientific) at any taxonomic rank.
        in: query
        required: false
        schema:
          type: string
      - name: filename
        description: Output file name.
        in: query
        required: false
        schema:
          type: string
          default: ncbi_dataset.zip
  /protein/accession/download:
    post:
      summary: Get a prokaryote gene data package by RefSeq protein accession
      description: Download a prokaryote gene data package including sequence, annotation and data reports by RefSeq non-redundant protein accession.
      tags:
      - Prokaryote
      operationId: download_prokaryote_gene_package_post
      responses:
        default:
          description: An unexpected error response.
          content:
            text/plain:
              schema:
                $ref: '#/components/schemas/rpcStatus'
        '200':
          description: A successful response
          content:
            application/zip:
              schema:
                format: binary
                type: string
                description: Zip compressed stream
      requestBody:
        required: true
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/v2ProkaryoteGeneRequest'
            examples:
              Multiple Prokaryote (WP_) accessions example:
                description: DNA-deoxyinosine glycosylase (WP_015878339.1) and type II CRISPR RNA-guided endonuclease Cas9 (WP_001040085.1), with gene and protein sequence files
                value:
                  accessions:
                  - WP_015878339.1
                  - WP_001040085.1
                  include_annotation_type:
                  - FASTA_GENE
                  - FASTA_PROTEIN
      parameters:
      - name: filename
        description: Output file name.
        in: query
        required: false
        schema:
          type: string
          default: ncbi_dataset.zip
components:
  schemas:
    v2ProkaryoteGeneRequestGeneFlankConfig:
      type: object
      properties:
        length:
          type: integer
    rpcStatus:
      type: object
      properties:
        code:
          type: integer
          format: int32
        message:
          type: string
        details:
          type: array
          items:
            $ref: '#/components/schemas/protobufAny'
    v2Fasta:
      type: string
      enum:
      - FASTA_UNSPECIFIED
      - FASTA_GENE
      - FASTA_RNA
      - FASTA_PROTEIN
      - FASTA_GENE_FLANK
      - FASTA_CDS
      - FASTA_5P_UTR
      - FASTA_3P_UTR
      default: FASTA_UNSPECIFIED
    v2ProkaryoteGeneRequest:
      type: object
      properties:
        accessions:
          type: array
          items:
            type: string
        include_annotation_type:
          type: array
          items:
            $ref: '#/components/schemas/v2Fasta'
        gene_flank_config:
          $ref: '#/components/schemas/v2ProkaryoteGeneRequestGeneFlankConfig'
        taxon:
          type: string
          title: Limit gene sequences and the genome assemblies described by the [annotation report file](https://www.ncbi.nlm.nih.gov/datasets/docs/v2/reference-docs/data-reports/prokaryote-gene-location/) to the specified NCBI Taxonomy ID or name (common or scientific) at any taxonomic rank.
    protobufAny:
      type: object
      properties:
        type_url:
          type: string
        value:
          type: string
          format: byte
  securitySchemes:
    ApiKeyAuth:
      type: apiKey
      in: query
      name: api_key
    ApiKeyAuthHeader:
      type: apiKey
      in: header
      name: api-key
externalDocs:
  description: Interactively explore genome assembly datasets
  url: https://www.ncbi.nlm.nih.gov/datasets/genomes/