How to choose k-mer for the genome size estimation?

Question

[524730309]

Hi, there. I am using Jellyfish-2 and Genomescope2 to estimate the genome size of my object. I choose different k-mers (e.g. 17 and 21) and they turned out very different. The same sample shows different genome size, for example when I am using 17 mer, the genome size is around 154379114 bp, but when I use 21 for test, the genome size is about 313880765 bp. So, here is my question:

1. How should I choose k-mer for the genome size estimation ? (I have already chosen the k according to the formula: 4^k > Genome Size210)
2. The data I am using is from the Illumina short reads, pair-ended. So should I using Trinity fastool to reverse one of fastq.gz and merge them together before k-mer analysis ?

I'll really appreaciate it if somebody could give me a hint.

[lieven.sterck]

out of interest, where did you came up with this formula: 4^k > Genome Size210 ?

Answer 1

[Oumaima]

I have the same question , how should I choose Kmer for the genome size ???? plz