On Friday 15 April 2011 15:33:04 Marco Fondi wrote:
> I have tried several times and changing a little bit the
> command line parameters but I got always the same result.
Well, you could have posted the command line you used to see whether
there were any more or less hidden traps (-SB:abnc comes to mind).
> Now, I want to go
> further with this assembly, trying to map the Illumina reads on the
> scaffolds and having extra contig from Illumina reads that cannot be
> mapped on the scaffolds. Everything is ok at the beginning but, in a few
> hours, all the space on my disk (almost 80G) is over and, of course, the
> assembly stops.
You also could have given some more specific information (how long are
the Solexas, how many contigs and how many megabases are the scaffold
sequences etc.). This usually helps to get things into perspective.
That being said: 80 GiB for a *mapping* assembly seems hefty ...
especially for just 5m Solexa reads. If you were doing de-novo with
that data set with anything less than the current development version,
I'd be tempted to say that in some unlucky cases you might get this
amount of log/tmp files.
> Do you have some idea on how to fix this?
Could you please try the current development version (3.2.1.13) ... it
contains quite some new things which were developed to reduce the disk
space used and get de-novo possible for 20 to 50m Solexa reads.
> Is this a known issue and I simply need more disk
> space available for MiRA to work?
More space (be it RAM or disk) is never wrong :-)
B.
