Hello Bremen,

We cannot offer advice about roast.

But regarding sorting, in our process it is taken care of at the 
chaining step.  The net step makes a single-coverage set of alignments 
which I believe is also important for multiz -- and definitely for 
phastCons.  Since you are not chaining and netting, our processing is 
different to achieve input for multiz/phastCons.

Examining the makedoc for hg19 could help. It has all of the exact 
commands we used to generate the Conservation track. This is in the kent 
source tree at:
kent/src/hg/makeDb/doc/hg19.txt
http://genome.ucsc.edu/FAQ/FAQdownloads.html#download27

A good place to ask for tool help might be the team that hosts Galaxy 
(PSU) to see if there is another way to ensure a single-coverage MAF (if 
you are not already filtering for single-coverage after sorting).
http://g2.bx.psu.edu/
[email protected]

Jennifer



---------------------------------
Jennifer Jackson
UCSC Genome Bioinformatics Group
http://genome.ucsc.edu/

On 2/22/10 10:14 AM, Bremen Braun wrote:
> Hi all,
> I'm having trouble getting my alignments recognized with PhastCons. I
> generated n alignments using lastz, did some filtering steps, and then used
> multiz n-1 times to get a file which are queries aligned to the same target.
> I generated a phylogenic tree with PhyloFit and this worked, but when I fed
> this all to phastCons, I'd get the following error:
> warning: maf_read: MAF file must be sorted with respect to reference
> sequence if store_order=TRUE.  Ignoring out-of-order blocks
> maf.c:323: failed assertion `idx_offset>= 0'
> Abort trap
>
> I used maf_order on the file taken as input to phastCons but I still got
> this error. I read somewhere that autoMZ can be used to generate a multiple
> alignment that respects phylogenic tree info but I am unable to see sample
> usage and the parameters are very vague. It looks like roast takes the same
> kind of parameters in the same order as autoMZ and I know that tool has been
> used for your data. Could you give me an example of how it's used?
>
_______________________________________________
Genome maillist  -  [email protected]
https://lists.soe.ucsc.edu/mailman/listinfo/genome

Reply via email to