velvet assembly problem with parameters

Question

[ilario.ferrocino]

Hi there I'm try to do assembly with velvet on my metagenomic illumina single end sequence.

I had try to use different kmer value but I have bad results like this:

How I can fix it?

parameters -short              sample name     total number contigs     total contig length     largest contig length     mean/median contig length     N50 (bp)
assembly kmer 53 e ins 300     09.TO           189,412                  146,667,097             54,723                    774/493                       914
assembly kmer 63 e ins 300     09.TO           178,173                  132,535,391             85,044                    744/470                       832
assembly kmer 82 e ins 300     09.TO           146,108                  103,053,100             177,456                   705/436                       746
assembly kmer 77 e ins 300     09.TO           155,024                  110,414,760             151,465                   712/442                       756
assembly kmer 60 e ins 300     09.TO           182,064                  138,638,656             63,991                    761/478                       873
assembly kmer 50 e ins 300     09.TO           194.84                   152,368,820             51,381                    782/501                       932

[Ram]

How did you conclude these results are bad? How do you think better results should look like? These inputs will help set some context.

Answer 1

[dago]

Although I think it would be nice if you specify what you want to improve, I think you can take a look at this program: VelvetOptimiser

Maybe you could also try different assembly strategy, in this paper there might be something interesting:

• Use of simulated data sets to evaluate the fidelity of metagenomic processing methods
• Evaluation of short read metagenomic assembly
• Sequence assembly demystified