Does velvet generate different assemblies from the same input data and same velvet parameters?

11 November 2013 3 6K Report

I am using four different libraries of read size 76bps, insert sizes are 300bps, 1kb, 8kb and 12kb. Expected genome size is 80MB.

I am running velvet using these four libraries. Actually I tried velvet first for different k-mers and then for the best k-mer I tried different cov_cutoffs. In all these assemblies from same k-mer (69) with different cov_cutoffs, I used the same Roadmap and Sequences files from the initial velveth run (K-mer 69 and cov_cutoff default).

Surprisingly I got 10MB of N50 and 23MB of largest scaffold size using cov_cutoff of 12 (median coverage is 30.76 in Log file) on the previously generated Roadmap and Sequences files. Then later I tried all new assembly from the same reads, K-mer 69 and cov_cutoff 12, now my N50 is 2MB and largest scaffold size is 6.78 MB.

Later I tried the same input files and same velvet parameters and I figured out that velveth is generating different Roadmap file for the same k-mer 69, for all three runs. What could be the reason behind this? In this case it is not possible to regenerate the results.

My question is, does velvet generates different assemblies from the same data and with same parameters?

I would really appreciate your response on this.

Sébastien Wielgoss

Hi Rahul.

Thanks for this post.

I have used Velvet for the assembly of a bacterial genome with only a 10th of the size of your genome (9Mb), but I also get different assembly results for each run on the same data. However, they do not vary dramatically as in your case. My approach also involves the inspection of the various assembly runs agaisnt the genome of an existing reference strain. If the assembly appears to be good I'll stick with it, if there are apparent artefacts I go with a different assembly run or assembler.

Specifics: I only used a single library prep of 500bps for an Illumina HiSeq 2000 machine, which produced reads of 101 bp lengths (paired end). The total number of clean reads approach 10 million (expexcted coverage of 100x for the genome; total data volume per genome = 1Gbp). For finding the optimal kmer size for the run I used the program VelvetK, which kmer I used for the velveth step.

And I ran velvetg like that:

velvetg /someplace/somefolder/ -ins_length 500 -exp_cov 100 -min_contig_lgth 500 -cov_cutoff auto

With kind regards

Sébastien

Rahul Sharma

Dear Sébastien,

Thanks for your reply. Ya I tested many times, velveth generates different Roadmap file for the same data. Daniel (Guy who wrote velvet) told me that, if we compile velvet with OpenMP it will generate different results, but if there is much difference in assembly parameters then our assembly is not reliable.

Thanks for your input.

Regards,

Rahul

Rahul Sharma

Dear Rohan,

Thanks for your reply. But I trimmed adapter/primers and filtered reads with Ns and reads below average Q26 at the early stages of data analysis. That could not be the issue. The variation is because of OpenMP usage.

Regards,

Rahul

Enhanced Yellow Fluorescent Protein (Aequorea victoria) DNA sequence?

Is it possible for two human neuronal cell lines to have different gene sequence for the same protein?

Runs of nucleotides in Sanger sequencing?

Has anyone used the Illumina ITS1 primer sets?

How do you mesh a thin solid in Fluent?

Errors in DNA sequences?

Is it possible to define biogas as a fuel in the reformer process(scwg) in Homer software?

I can detect my protein but the encoding DNA sequence has stop codons within. Should I ignore the stop codons?

What is the significance of (isolated) high titers of IgG anti GD1a antibodies in an older patient with polyneuropathy?

Repository for capture-seq raw reads?