Phix genome kmer length

Question

[Bioblazer]

I am working on very small genome size 5500 bp. For de novo assembly how to pick optimal kmer for this genome size.

[Brian Bushnell]

It's not clear what you are trying to accomplish, or what kind of data you have, so there's no way to give a good answer to this question. Can you please give a complete description of the situation?

[Bioblazer]

I have Miseq sequencing data for phage virus, the genome size about 5kb. For de novo assembly,depending upon kmer, our contig assemblies will vary. I want to know what is ideal kmer value or range of kmer values for 5kb genome size.

Answer 1

[Brian Bushnell]

Assembly should take seconds, so the best option is to use all kmer lengths up to ~95% of read length and pick the best assembly. In this case, the best assembly will be any that produce a single ~5500bp contig. If you are having trouble producing a good assembly, I suggest trying Tadpole in the BBMap package, which tends to do a good job assembling viruses that for whatever reason produce strange results in other assemblers such as Spades.