Mira assembler: Medium sized genomes;How to use 2 separate files for paired-end reads

[ndeshpan]

Hi all,

This is Nandan Deshpande from UNSW, Sydney, Australia.

I am working trying to assemble a eukaryotic genome of estimated size of around 400 MB.
I have used SOAPdenovo and produced a rough assembly. As a comparison I am trying to use MIRA assembler. I have few questions:


I have 3 paired-end read data-sets from Illumina Solexa (102 bp ;30917380 reads per file)
With a RAM memory of around 200+ GB can anyone suggest if MIRA has the capacity to assemble these reads? Does it use multiple processors?


Also ..
I am finding difficult to exactly understand the command for MIRA; the nomenclature for the files etc..Is it essential that I have the paired-end reads for a pair in a single file? I have them in two separate files at this time (containing 0/1 and 0/2 reads respectively)..Can anyone who has worked with mira and paired-end reads please paste a sample command with my above requirements?

Does mira help in scaffolding using only paired end reads (insert sizes avg 300bp); but no mate-pairs?

Appreciate all responses..

cheers,

Nandan

[tonybolger]

I have 3 paired-end read data-sets from Illumina Solexa (102 bp ;30917380 reads per file)
With a RAM memory of around 200+ GB can anyone suggest if MIRA has the capacity to assemble these reads? Does it use multiple processors?

Our experience trying to assemble 50GB of 454 data with MIRA suggests you won't want to wait that long.

If it's just solexa data, i'd suggest keeping to the various de-bruijn assemblers - not sure i can really 'recommend' one though, i've had bad experiences with all of them that i tried.

BTW, SOAP doesn't appear use paired data until the scaffolding stage, so it's not necessarily as bad as it looks.

[sphil]

Hey,

first of all you are right. Due to massive amount of flags mira is using its hard to come up with. Here, http://mira-assembler.sourceforge.ne...e_library_size, you can find a very nice description on how to assemble paired end reads. Nevertheless i would recommend to subscribe at mira mailing list. The are really fast and even Bastien tries to help as much as he can (if the manual doesn't work out)...
Only havin' a slight look i would say you need to concate your files into one.
Then just start assembly with:
mira
--project=XXXXX
--job=denovo,genome,accurate,solexa
SOLEXA_SETTINGS -GE:tismin=250:tismax=750
>&log_assembly.txt

accurate means that you will have to wait until sun burns out so i would change it to draft for a first shot.

SOLEXA_SETTINGS -GE:tismin=250:tismax=750 , means that your insert size is 500bp.... everything else might be quite ok with defualt solexa settings

best,

philip

[ndeshpan]

thanks guys