[mira_talk] Re: Mira Question: Solexa reads...memory error or bug?
- From: Bastien Chevreux <bach@xxxxxxxxxxxx>
- To: mira_talk@xxxxxxxxxxxxx
- Date: Fri, 13 Mar 2009 23:09:05 +0100
On Friday 13 March 2009 Alex Washington wrote:
> Greetings, I have throughly searched through the Mira package for an answer
> to this problem but to no avail.
> [...]
Hello Alex,
the manuals for Solexa could be ... somewhat more extensive, I agree :-)
However, your problem seems to be something peculiar. Let's start with the
side note (which are probably not interesting to you but might be important to
me):
1) Could you please send me the FASTA file you are loading the backbone from?
There's something strange in the log as MIRA says it loaded two sequences but
just added rails for one backbone. This is not normal.
2) It looks like your backbone is the sequence from a GenBank entry. For
prokaryotic entries, MIRA has no problem loading a GenBank entry directly,
including the complete annotations ... which opens up a some neat goodies
further down the analysis pipeline. You might want to try that some day.
Now to your out of memory problem: it comes at a very surprising point, just
when MIRA allocates space to get the sequences loaded.
Questions:
1) Would you be running a 32 bit executable? Then, even if your system has
dozens of gigabytes, a process can use only 2.7 to 3.5 GiB. Which might lead
to out of memory.
2) If you are running a 64 bit executable: what's the system memory and how
many reads are in the Solexa file?
Now, before answering that yourself: could you please download
http://www.chevreux.org/tmp/mira_2.9.42x1_dev_linux-gnu_x86_64.tar.bz2
and send me the output of that version? I've added the needed things to the
log so I can look it up myself.
Besides, in 2.9.42 I've activated new Solexa clipping routines which are way
faster and better than the old ones. With this version I assembled and
performed mutation analyses on 5 Solexa projects on Wednesday and Thursday, so
it's quite well tested.
Regards,
Bastien
--
You have received this mail because you are subscribed to the mira_talk mailing
list. For information on how to subscribe or unsubscribe, please visit
http://www.chevreux.org/mira_mailinglists.html
Other related posts: