iCORN2 corrects PacBio assemblies

[ebioman]

The assembly is actually a classical Illumina type and PacBio reads are used purely for scaffolding analysis.

I will give it a try next week, THX

[ThomasDOtto]

Well, it might scale ok. All the components are working fine on large genome, the mapper (SMALT), GATK SNP, and the core icorn should work also. The only thing that won't work would be the checking of the perfect mapping reads. But we could twist the variant calling to be more stringent, so one could skip that.

On another note, now is the assembly coming together? 1G is pretty big for PacBio...

[ebioman]

I was afraid of that reply, since I am currently looking at a genome with ~1G and Illumina reads with 100x coverage
I might give it at try for the final scaffolds nevertheless. Do you have any experience with data that large for your program ? I am wondering since I discovered that most PacBio tools are actually pretty limited to Bacteria genome sizes..

[ThomasDOtto]

Icorn2 works on assembled genomes. It is quite powerfull to correct frameshifts that quiver does not correct.
On a 20MB genome, with three iterations and around 100x Illumina coverage it should take around a day. To correct a 4MB bacteria takes 1h per iteration.

--Thomas

[ebioman]

Sounds good to have an alternative !
Can you give some estimation how long it takes to correct e.g. one PacBio cell with your program ?

[ThomasDOtto]

We released iCORN2:

http://icorn.sourceforge.net/

iCORN2 was applied to correct several PacBio assemblies (Bacteria / Plasmodium). It requires short read pairs, and iteratively corrects errors, which in PacBio mostly indels.

http://bioinformatics.oxfordjournals.org/content/26/14/1704.short

The current tar ball can be found at
ftp://ftp.sanger.ac.uk/pub4/resource...orn2.V0.95.tgz

Happy about (any) feedback!

--Thomas