optimal k value?

[LHT]

Hi everyone, my question:
How to determine the optimal k value (in assembling using k-mer)?

The k-mer length we used includes 40, 50, 60 and 70. We have the number of clusters and singletons produced. How do I know which one is the best to choose to proceed with assembly with MIRA?

I am new in bioinformatics, please let me know if the information I've given is not clear. Thanks for your help!

[LHT]

To add, my data is 454 reads.
Any helpful comment is appreciated.

[biomed]

In velvet you can use optimizer or assemblyassember tools included in velvet package...

[Himalaya]

Hi everyone, my question:
How to determine the optimal k value (in assembling using k-mer)?

The k-mer length we used includes 40, 50, 60 and 70. We have the number of clusters and singletons produced. How do I know which one is the best to choose to proceed with assembly with MIRA?

I am new in bioinformatics, please let me know if the information I've given is not clear. Thanks for your help!

I don't think that we can say what is the optimal k value. It depends on the read lengths of all the reads, if i am not wrong. i tried velvet with a simulated reads rather than actual reads generated from sequencer. i increased the reads from 21 to 57. the kmer 57 gave me the highest n50. kmer61 gave the n50 lower than kmer 57 so i didnt increase the kmer size again.
Hope this reply makes u happy

[lmf_bill]

I remember abyss suggests using k value loop to find the proper one.