Kmergenie histogram file

Question

[Rahul]

---

Hello,

I have used Kmergenie for determining the Kmer value for my data.I got 41 as best Kmer value using default parameters.

I am not sure about this value.Can anybody help me to sort out this problem?.

Following is the link for my kmergenie result containing histogram file.

http://dropcanvas.com/7j78z440zGeZ1e

Answer 1

[apelin20]

kmergenies tells you which k-mer size results into maximum genomic unique k-mers. Did you try assembling?

[Rahul]

Thank you very much for your reply.

I have no previous experience of this work.I have not started assembling till now,just done up to velvet H and velvet G with Kmer value 150 calculated without using kmergenie and velvetK.

At the presentt now I am not sure about my kmergenie histogram,though I got best Kmer value 41.I bit confuse weather to opt 41 or 82 as best kmer value beacuse there are more than 1 pick in histogram or shall I used velvetk.

[apelin20]

what is your nucleotide coverage and what are your read lengths?

[Rahul]

Partial genomic fosmid library

Fastqc report

FORWARD SEQUENCE

Total sequence 522196
Sequence length 151

REVERSE SEQUENCE

Total sequence 522196
Sequence length 151

[apelin20]

I can suggest you try SPAdes, it utilizes a multi k-mer approach, do you can use 41,51,61,81....

http://bioinf.spbau.ru/spades

spades.py -k 21,33,55,77 --careful --only-assembler <your reads> -o spades_output

[Rahul]

Thanks apelin,

I will use SPAdes as you suggested,but for your kind information I used Kmergenie with Kmer value range in between 15 to 121.K mergenie generated lots of histogram for different value out of that,I got 41 as best Kmer value.

Right now I am following one tutorial protocol for doing my analysis.The protocol link I have mentioned below.Can I get your suggestion on this protocol,I think it would be very helpful for me.

http://www.microbialinformaticsj.com/content/supplementary/2042-5783-3-2-s1.pdf

Thanks

[Rayan Chikhi]

Hi! Kmergenie dev here. The advice given in this response is sound. Kmergenie is only useful when you want to perform assembly, and when your assembler (e.g. Velvet) asks for a single k value as a parameter. SPAdes is an excellent assembler for bacterial genomes (it generally outperforms Velvet), and does not even require to pass a k value as parameter, as it uses a multi-k approach. So, if you goal is to perform bacterial genome assembly, you could simply skip Kmergenie and run SPAdes directly. For other cases (e.g. eukaryotic genome, where SPAdes doesn't apply), then Kmergenie can be combined with a single-k assembler (Velvet, ABySS, Ray, SOAPdenovo, Minia, etc..).

[Rahul]

Thanks Rayan for showing much interest in my post

For your kind information I am working with plant genome sequence (partial library).....not with bacterial genome.I am very new to this field only since last month I am working on this topic.I have generated assembly file as contigs.fa using velvet with k mer size 41. my contigs.fa file contain NNNNNN bases.Does it mean that file is a scaffold or only assembly file?

Can I use velvet for scaffolding purpose?

Thanks in advance

[Rayan Chikhi]

It's a scaffold. Velvet's contigs.fa file is a notorious misnomer :) whenever you have paired-end/mate-pair data, it will contain scaffolds.

[Rahul]

Right now I am trying to use Opera scaffolder, but getting failed.I would be very grateful if you could suggest me any scaffolder .

[Rayan Chikhi]

SSPACE is quite easy to use; or BESST (which has a manual here)

[Rahul]

Thanks Rayan I am using SSPACE. As you have suggested I will also try BESST.

Cheers
Rahul