[Genome] blastz-normalizeLav, fasta-subseq, strip_rpts etc. details sought]

[Hiram Clawson]

Good Afternoon Malcolm:

I've been corrected on my answer to you:

> Sorry Hiram, this is not quite right... fasta-subseq and strip_rpts are still used (but only for
> abridging of repeats). They are called by src/utils/blastz-run-ucsc. We use execs compiled
> from PSU source.
>
> partitionSequence takes the chrom.sizes info and splits large sequences into chunks
> and/or lumps small sequences together, to meet the target sizes given in DEF.
> Actually, partitionSequence doesn't even do the splitting, only the specification
> of the splitting. blastz-run-ucsc uses the specification to extract sequence (from 2bit, nib or fasta).

Hiram Clawson wrote:
> Perhaps the track documentation is a bit out of date.
>
> The ancient fasta-subseq and strip_rpts scripts have been
> rewritten into the single script: src/utils/partitionSequence.pl
> which does nothing more than take repeat masker information
> and cuts those sequences out of the input to avoid matching
> repeats between two genomes.
>
> Our entire blastz process is now run in an automated fashion with
> the single script: src/utils/doBlastzChainNet.pl which usually
> runs for 24 to 48 hours with several kilokluster jobs to produce
> all the chain and net data for those tracks.
>
> --Hiram

[Cook, Malcolm]

Hiram,

Thanks for the additional explanations.

So, 1,000 PIIIs for 1-2 days, eh? OK - I get the picture... do not try
this 'at home'!

Can you or someone tell me how you decide which genomes to produce a
'self alignment' analysis? I see you have Dog, Human, Mouse, Zebrafish.
Might there be ever a plan to produce one for D. melanogaster?

Thanks,

--Malcolm

[Hiram Clawson]

Good Morning Malcolm:

We do self alignments for the important vertebrate genomes.

D. melanogaster is *only* 133,000,000 bases.
Perhaps it could be done 'at home'
13 chroms against 13 chroms would be only 169 blastz jobs.

--Hiram