On May 12, 2009, at 3:08 AM, michel bellis wrote:

>
> Chris,
>
> The results I have sent are in order as I entered in the pydev console
> (eclipse on Windows XP). So the os.listdir is before construction of
> the NLMSA.

that's odd.  Why would there be .idDict and .seqIDdict files *before*  
building the NLMSA.  I guess you already tried to build it before, and  
these are left over from old builds.

>
>
> I tried your code and got a lot of file longer than 1 000 000. But I
> do not undestand what is the limit exactly.

The limit is that my brain was inoperative when I wrote that.  I meant  
1 billion, because I wanted it to report any sequence that might go  
over the 32 bit int limit.  But what's 3 orders of magnitude between  
friends, anyway?

>
> I ran it with a different value (>= 2 000 000 000) and was surprised
> to find this:
> mm8.chrM 2666011489
>
> So I reconstructed mm8 without chrM. And then at the end I got this:
> mm8.chr9_random 2666012422 (which was hg18.chr9_random 1146434

OK, as I thought, something is going wrong with the reading of your  
sequence file. Was chrM the *last* sequence in the file?  And  
chr9_random the next-to-last sequence?  That would imply that the last  
sequence length is always getting screwed up.  Looking at the  
seqfmt.pyx code, I don't see any obvious reason why that could occur.   
Assuming that 'mm8' is the name of your fasta input file, try the  
following test code:

from pygr import seqfmt
f = file('mm8')
d = {}
seqfmt.read_fasta_lengths(d, f, 'mm8')
print d # see if the sequence length is already crazy at this point

Finally, what platform (Windows via Cygwin?) and Python / Pyrex  
versions are you using?

-- Chris

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"pygr-dev" group.
To post to this group, send email to pygr-dev@googlegroups.com
To unsubscribe from this group, send email to 
pygr-dev+unsubscr...@googlegroups.com
For more options, visit this group at 
http://groups.google.com/group/pygr-dev?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to