Reference genomes need a new abstractions

[ihodes]

We hard-code in things like the Ensembl release version into a reference genome like mm10 that shouldn't be there; the release isn't tied to a given genome. This probably applies to things like the dbsnp, exnome_gtf, as well (cc @arahuja)

type t = private {
    name : name;
    ensembl : int;
    species : string;
    metadata : string option;
    fasta : Location.t;
    dbsnp : Location.t option;
    known_indels : Location.t option;
    cosmic : Location.t option;
    exome_gtf : Location.t option;
    cdna : Location.t option;
    whess : Location.t option;
    major_contigs : string list option;
  }

[smondet]

This is not a bug?

A Reference_genome.t is a fixed collection of reference-genome-related things working together.

If you want b37 with ensembl 59 and a custom COSMIC, you can just create it, and add it to your Biokepi.Machine.t.

[iskandr]

Some of those fields never change (species, major_contigs), some change
between patches of a reference (e.g. fasta), and others change on every
Ensembl release (e.g. exome_gtf, cdna).

On Tue, Sep 27, 2016 at 1:39 PM, Sebastien Mondet [email protected]
wrote:

This is not a bug?

A Reference_genome.t is a collection of reference-genome-related things
working together.

If you want b37 with ensembl 59 and a custom COSMIC, you can just create
it, and add it to your Biokepi.Machine.t.

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
#365 (comment),
or mute the thread
https://github.com/notifications/unsubscribe-auth/AAC9OdULmKgBx2Bn-lXjrXpvFA-67xU6ks5quVTlgaJpZM4KH77i
.

[ihodes]

Not a bug, that was a mistaken label, but I do think something needs to change in the code, not just the docs (at the very least the name; calling this collection of things a Reference_genome is misleading).