[mira_talk] Re: Memory problem with MIRA

Hi Bastien,

These are EST reads. Following is the ls -l output of the directory

total 960M
drwxr-xr-x 2 ankitgpt users    6 2009-12-03 13:42 1TCA_d_chkpt
drwxr-xr-x 2 ankitgpt users   41 2009-12-03 13:42 1TCA_d_info
drwxr-xr-x 2 ankitgpt users 4.0K 2009-12-03 13:50 1TCA_d_log
drwxr-xr-x 2 ankitgpt users    6 2009-12-03 13:42 1TCA_d_results
-rw-r--r-- 1 ankitgpt users 246M 2009-12-03 01:40 1TCA_in.454.fasta
-rw-r--r-- 1 ankitgpt users 714M 2009-12-03 01:43 1TCA_in.454.fasta.qual
-rw-r--r-- 1 ankitgpt users  37K 2009-12-03 13:57 log_assembly.txt
-rw------- 1 ankitgpt users  37K 2009-12-03 12:55 nohup.out

I was using --noclipping because of the note in the usage manual that said
the minimum coverages for 454 should be around 14-15x. I have an average
coverage of 2x it seems in my project. So I didn't want to loose any
information. Besides all the input sequences to mira have been processed
using Seqclean and Repeatmasker softwares.

I am using the OpenSuse 11.1 operating system and I guess its a 32 bit
system. I was able to get the miraMEM working and it puts an estimate at 4
GB of RAM for 650000 GS FLX Titanium reads. I'll download the latest version
and send you the first 50 lines as you had requested.

So do I need to get a 64 bit operating system to run it????

Thanks
Ankit Gupta
Virginia Tech

On Thu, Dec 3, 2009 at 3:25 PM, Bastien Chevreux <bach@xxxxxxxxxxxx> wrote:

> On Donnerstag 03 Dezember 2009 ankit gupta wrote:
> > I am trying to use mira for a de novo assembly of a transcriptome but
> after
> > running for a few minutes, it gives me a memory related error.
> >
> >      Writing temporary hstat files:
> >      [0%] ....|.... [10%] ....|.... [20%] ....|.... [30%] ....|.... [40%]
> > ....|.... [50%] ....|.... [60%] ....|.... [70%] ....|.... [80%] ....|....
> > [90%] ....|.... [100%] done
> >
> >      Analysing hstat files:
> >      [0%] ....|.... [10%] ....|.... [20%] ....|.... [30%] ....|.... [40%]
> > ....|.... [50%] ....|.... [60%] ....|.... [70%] ....|.... [80%] ....|....
> > [90%] ....|.... [100%]
> >      Ouch, out of memory detected.
> >
> > My data is from 454 GS FLX Titanium series and I have nearly 650,000
> reads
> > with associated quality file. I am using a strong server with 32 gb RAM.
> >  Any idea why this might be happening?
>
> Hello Ankit,
>
> ummmm ... it throws an out of memory during the analysis of the hashes?
> This
> is ... peculiar. Am I right in guessing that these are not genomic reads
> but
> EST?
>
> Can you please give me an "ls -l" of the log directory, I'd like to check
> the
> size of the files.
>
> > I am using the following command to run mira:
> >      mira -project=1TCA --job=denovo,est,normal,454 -fasta -notraceinfo
> > -noclipping >&log_assembly.txt
>
> Oh, yeah, right, ESTs. There it is.
>
> Well, you are using "--noclipping" to kill off all the clips. Why? At least
> the Poly-A/Poly-T clip should be on in EST assemblies, else this will lead
> you
> nowehere (or into memory problems :-)
>
> > Also I am unable to figure out how to invoke the miraMEM memory call.
> When
> >  I try giving this as command, my shell returns with command not found
> and
> >  mira doesn't seem to accept is as an option or argument. Can you tell me
> >  what linux command should I give to invoke miraMEM to get an estimate of
> >  how much memory my project might need ?
>
> 'miramem'. If you downloaded the binary distribution, you'll see it in the
> 'bin' directory, realised as link to 'mira'. If you compiled yourself:
> create
> a symbolic link from mira to miramem (ln -s mira miramem).
>
> 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: