esembl compara perl api help needed!

[svl]

Hi everyone,

I'm trying to work with the ensembl perl api, but seem to be unable to do what I want. API -> http://www.ensembl.org/info/data/api.html

What I want is this:
I have a set of genomic positions from snp-calling. For each of the positions I want to know how conserved these single positions are compared to/in other species. So I figured I use the Compara API, which contains whole genome alignments, and some of the code here: http://www.ensembl.org/info/docs/api..._tutorial.html So basically I am looking for something like:

$bases_for_pos_1 = $alignment_human2mouse->getBases($human_position1)
$bases_for_pos_2 = $alignment_human2mouse->getBases($human_position2)
etc...(simplified, not real compara code)

The problem:
I can retrieve a 'GenomicAlignBlock' which contains the position I am looking for, but also a whole lot of surrounding sequence and I wander how I can get just the bases (from human and the other species) at the specific position...? Or even just whether they are equal or not.

Anyone?

-Stef

[whsqwghlm]

You may be best posting on the ensembl-dev mailing list, but here is a guesstimate;

First you need to restrict the Bio::EnsEMBL::Compara::GenomicAlignBlock to the region that you are interested in, using the restrict_between_reference_positions method (see http://www.ensembl.org/info/docs/Pdo...ock.html#POD31)

You can then call the slice method on the returned GenomicAlignBlock to return the genomic region of the reference. The returned Bio::EnsEMBL::Slice has a seq method to pull the sequence. The non-reference(s) can be accessed in a similar way.

[svl]

using the restrict_between_reference_positions method

Thank you! Works like a charm. Although I still don't see why they didn't include a method to directly pull one base-pair from the alignment, it seems so useful to me But anyhow my code is a little less long and more straight forward now.

You can then call the slice method on the returned GenomicAlignBlock to return the genomic region of the reference

This is not even needed, there is a method "alignment_strings" that returns the sequences (in my case 1 letter long) as an array of strings directly.

thanks again!