Biokepi API : Biokepi_run_environment.Reference

struct

  module Location = struct

    type t = [

      | `Url of string

      | `Vcf_concat of (string * t) list (* name × location *)

      | `Concat of t list

      | `Gunzip of t (* Should this be guessed from the URL extension? *)

      | `Untar of t

    ]

    let url u = `Url u

    let vcf_concat l = `Vcf_concat l

    let concat l = `Concat l

    let gunzip l = `Gunzip l

    let untar l = `Untar l

  end



  type t = {

    name: string;

    metadata: string option;

    fasta: Location.t;

    dbsnp: Location.t option;

    cosmic: Location.t option;

    exome_gtf: Location.t option; (* maybe desrves a better name? *)

    cdna: Location.t option;

    major_contigs: string list option;

  }



  let create

      ?metadata

      ~fasta

      ?dbsnp

      ?cosmic

      ?exome_gtf

      ?cdna

      ?major_contigs

      name = {

    name;

    metadata;

    fasta;

    dbsnp;

    cosmic;

    exome_gtf;

    cdna;

    major_contigs;

  }



module Default = struct



  let major_contigs_b37 =

    List.init 22 (fun i -> sprintf "%d" (i + 1))

    @ ["X"; "Y"; "MT";]

  let major_contigs_hg_family =

    List.init 22 (fun i -> sprintf "chr%d" (i + 1))

    @ [

      "chrX";

      "chrY";

      "chrM";

    ]

  let major_contigs_mm10 =

    List.init 19 (fun i -> sprintf "%d" (i + 1))

    @ [ "X"; "Y" ]



  module Name = struct

    let b37 = "b37"

    let b37decoy = "b37decoy"

    let b38 = "b38"

    let hg18 = "hg18"

    let hg19 = "hg19"

    let mm10 = "mm10"

  end





  (* Used by both B37 and B37decoy *)

  let b37_dbsnp_url =

    "ftp://gsapubftp-anonymous@ftp.broadinstitute.org/bundle/2.8/b37/dbsnp_138.b37.vcf.gz"

  let b37_cosmic_url =

    "http://www.broadinstitute.org/cancer/cga/sites/default/files/data/tools/mutect/b37_cosmic_v54_120711.vcf"

  let b37_exome_gtf_url =

    "http://ftp.ensembl.org/pub/release-75/gtf/homo_sapiens/Homo_sapiens.GRCh37.75.gtf.gz"

  let b37_cdna_url =

    "http://ftp.ensembl.org/pub/release-75/fasta/homo_sapiens/cdna/Homo_sapiens.GRCh37.75.cdna.all.fa.gz"



  let b37 =

    create Name.b37

      ~metadata:"Provided by the Biokepi library"

      ~major_contigs:major_contigs_b37

      ~fasta:Location.(

          url "ftp://gsapubftp-anonymous@ftp.broadinstitute.org/bundle/2.8/b37/human_g1k_v37.fasta.gz"

          |> gunzip)

      ~dbsnp:Location.(url b37_dbsnp_url |> gunzip)

      (* Alternate?

         "ftp://ftp.ncbi.nlm.nih.gov/snp/organisms/human_9606/VCF/v4.0/00-All.vcf.gz"

      *)

      ~cosmic:Location.(url b37_cosmic_url)

      ~exome_gtf:Location.(url b37_exome_gtf_url |> gunzip)

      ~cdna:Location.(url b37_cdna_url |> gunzip)



  let b37decoy =

    create Name.b37decoy

      ~metadata:"Provided by the Biokepi library"

      ~major_contigs:major_contigs_b37

      ~fasta:Location.(

          url

            "ftp://ftp.1000genomes.ebi.ac.uk/vol1/ftp/technical/reference/phase2_reference_assembly_sequence/hs37d5.fa.gz"

          |> gunzip)

      ~dbsnp:Location.(url b37_dbsnp_url |> gunzip)

      ~exome_gtf:Location.(url b37_exome_gtf_url |> gunzip)

      ~cosmic:Location.(url b37_cosmic_url)

      ~cdna:Location.(url b37_cdna_url |> gunzip)



  let b38 =

    (* Release 79 *)

    let b38_url =

      "ftp://ftp.ensembl.org/pub/release-79/fasta/homo_sapiens/dna/Homo_sapiens.GRCh38.dna.primary_assembly.fa.gz" in

    let dbsnp_b38 =

      "http://ftp.ncbi.nlm.nih.gov/snp/organisms/human_9606_b142_GRCh38/VCF/00-All.vcf.gz" in

    let gtf_b38_url =

      "http://ftp.ensembl.org/pub/release-79/gtf/homo_sapiens/Homo_sapiens.GRCh38.79.gtf.gz" in

    let cdna_b38_url =

      "http://ftp.ensembl.org/pub/release-79/fasta/homo_sapiens/cdna/Homo_sapiens.GRCh38.cdna.all.fa.gz" in

    create Name.b38

      ~metadata:"Provided by the Biokepi library"

      ~major_contigs:major_contigs_b37

      ~fasta:Location.(url b38_url|> gunzip)

      ~dbsnp:Location.(url dbsnp_b38 |> gunzip)

      ~exome_gtf:Location.(url gtf_b38_url |> gunzip)

      ~cdna:Location.(url cdna_b38_url |> gunzip)



  let hg18 =

    let hg18_url =

      "ftp://gsapubftp-anonymous@ftp.broadinstitute.org/bundle/2.8/hg18/Homo_sapiens_assembly18.fasta.gz" in

    let dbsnp_hg18_url =

      "ftp://gsapubftp-anonymous@ftp.broadinstitute.org/bundle/2.8/hg18/dbsnp_138.hg18.vcf.gz" in

    create Name.hg18

      ~metadata:"Provided by the Biokepi library"

      ~major_contigs:major_contigs_hg_family

      ~fasta:Location.(url hg18_url|> gunzip)

      ~dbsnp:Location.(url dbsnp_hg18_url |> gunzip)

  let hg19 =

    let hg19_url =

      "ftp://gsapubftp-anonymous@ftp.broadinstitute.org/bundle/2.8/hg19/ucsc.hg19.fasta.gz" in

    let dbsnp_hg19_url =

      "ftp://gsapubftp-anonymous@ftp.broadinstitute.org/bundle/2.8/hg19/dbsnp_138.hg19.vcf.gz" in

    create Name.hg19

      ~metadata:"Provided by the Biokepi library"

      ~major_contigs:major_contigs_hg_family

      ~fasta:Location.(url hg19_url|> gunzip)

      ~dbsnp:Location.(url dbsnp_hg19_url |> gunzip)



  let mm10 =

    let mm10_url =

      "ftp://ftp.ensembl.org/pub/release-84/fasta/mus_musculus/dna/Mus_musculus.GRCm38.dna_sm.primary_assembly.fa.gz" in

    let dbsnp_mm10_snps_url =

      "ftp://ftp-mouse.sanger.ac.uk/REL-1303-SNPs_Indels-GRCm38/mgp.v3.snps.rsIDdbSNPv137.vcf.gz" in

    let dbsnp_mm10_indels_url =

      "ftp://ftp-mouse.sanger.ac.uk/REL-1303-SNPs_Indels-GRCm38/mgp.v3.indels.rsIDdbSNPv137.vcf.gz" in

    let gene_annotations_gtf =

      "ftp://ftp.ensembl.org/pub/release-79/gtf/mus_musculus/Mus_musculus.GRCm38.79.gtf.gz" in

    create Name.mm10

      ~metadata:"Provided by the Biokepi Library"

      ~major_contigs:major_contigs_mm10

      ~fasta:Location.(url mm10_url |> gunzip)

      ~dbsnp:Location.(

          vcf_concat ["db_snps.vcf", url dbsnp_mm10_snps_url |> gunzip;

                      "db_indels.vcf", url dbsnp_mm10_indels_url |> gunzip]

        )

      ~exome_gtf:Location.(url gene_annotations_gtf |> gunzip)



end



end