Velvet Raw Seq Assembler

Question

[tanjafiegel]

To assemble fastq raw seq read from Ion Torrent for bac genome (3Mb), what are the 'best' commands to get the optimum contig alignment on average (velveth and velvetg).

[Simon Cockell]

this might be a helpful link to get you started: http://dna.med.monash.edu.au/~torsten/velvet_advisor/

Answer 1

[Leszek]

You can try Newbler or Mira

[Lee Katz]

Newbler seems to work the best for me so far, for Ion Torrent. VelvetOptimiser worked OK. On a recent de novo assembly with Newbler I got 465 contigs, N50: 17kb, and an assembly size of 4.1 Mb. With VelvetOptimiser, I got 1480 contigs, 3.9 Mb, and N50: 4kb. When I merged with the combiner in CG-Pipeline (which uses Minimus2), I got a final assembly of 454 contigs, N50: 19.5 k, and a genome length of 4.2 Mb.

I have not tried Mira for Ion Torrent and so I cannot comment on that.

Answer 2

[Jeremy Leipzig]

i don't think velvet would be good for "homopolymer-ish" data. it is bad for 454 and it behaves even worse for ion torrent.

http://pathogenomics.bham.ac.uk/blog/2011/05/first-look-at-ion-torrent-data-de-novo-assembly/