hammerlab / ogene

Type-safe scripts for genomic file wrangling

What is this?

This is a small program that karyotypically sorts a fasta file.

That is, say you have a fasta file with contigs in lexicographic order, for example, the first contig is 1 (or chr1) and the second is 10, then 11, and so forth; GATK and other tools may not be too happy about this. You can solve this by running this tool:

fasta-orderer unsorted-fasta.fa sorted-fasta.fa 

You'll likely have to wait a couple minutes (~1 minute for a 3GB fasta on my computer), as the entire fasta is loaded into memory. Sorry.

Now, the contigs will be in order from 1, 2, ..., 22, ..., X, Y, MT, ... and so forth.

And that's all there is to it.

Building

You'll need opam (probably a brew install opam away from you, at most), and then you'll need to do the following:

make deps
make

Installing these dependencies might take a few minutes; for that, your forgiveness is begged.

From here, you should be good to go.

Contributing

Contributions welcome; please see the Github issues page.

Testing

You'll need to opam testDeps, then run the tests with make all; make test.

ogene's Issues

More efficiently handle the sequence splicing

Derive default sequence line length from input fasta

Default shouldn't be 60; should be what it was before.

Add tests

Needs to make sure we output the same fasta, just in a different order

Support alternate orderings

Should be able to sort in any specified order. Taking a .dict file would be a nice start. May as well support lexicographic ordering, too; you never know what'll get thrown at you.

What other utilities should go in here?

As an exercise, would be nice to have a flagstat command.

Avoid holding the whole fasta in memory

Totally unnecessary.

Add to opam

Use HTTPS repo

https://github.com/hammerlab/ogene/blob/master/opam/opam#L8

Don't discard headers in fasta

AFAIK I can't get the headers back out of Biocaml.Fasta.headers, but I should be able to/hack around this if so.

Fix opam file

Something is not formatted correctly here: https://github.com/hammerlab/ogene/blob/master/opam/opam

             package: ogene
             version: 0.0.0
              pinned: git (d5f12b2d)
        upstream-url: https://github.com/hammerlab/ogene.git
       upstream-kind: git
            homepage: https://github.com/hammerlab/ogene
         bug-reports: https://github.com/hammerlab/ogene/issues
            dev-repo: [email protected]/hammerlab/ogene.git
              author: Isaac Hodes <[email protected]>
             license: Apache-2.0
   installed-version:
   available-version: 0.0.0
         description:

Not showing dependencies, and opam install isn't installing dependencies.