k-mer optimization in Trinity

Question

[seta]

Hi all,

As far as I know the k-mer in the trinity software is almost constant (25), I would like to know if it should be changed and optimized to make the best transcriptome assembly from about 400 million reads. any suggestion would be highly appreciated.

Answer 1

[snek]

I found somewhere that the kmer can be changed between 25 and 32 for trinity. Is it possible and advisable to increase it to 75-99 for assembly of illumina reads with 150 bases?

[Rohit]

It definitely is advisable, but not sure if possible. At least the earlier versions of trinity (2-3 years ago) had fixed kmer sizes in chrysalis step (clustering), but other steps can use a user-defined kmer. Additionally, there was an upper bound of 32 for kmer length.