Putting it up front: what is the optimal Kmer setting for a given set of NGS reads?
I 'know', there is Velvet Optimizer and that one has to test the best Kmer setting empirically, but:
when using different assemblers with different kmers to assemble a genome (50 to 400 Mb size range) without knowing the actual genome size (hope to measure it soon) you end up highly confused. Well, we have good looking assemblies with quite different genome sizes depending on the Kmer setting. Recently I heard that one should chose a Kmer at least half as large as ones read size (say 50 for 100bp Illumina pe), what about that one? I would think that one loses a lot of information (sensitivity, is it?) that way.
Ciao
Phil
I 'know', there is Velvet Optimizer and that one has to test the best Kmer setting empirically, but:
when using different assemblers with different kmers to assemble a genome (50 to 400 Mb size range) without knowing the actual genome size (hope to measure it soon) you end up highly confused. Well, we have good looking assemblies with quite different genome sizes depending on the Kmer setting. Recently I heard that one should chose a Kmer at least half as large as ones read size (say 50 for 100bp Illumina pe), what about that one? I would think that one loses a lot of information (sensitivity, is it?) that way.
Ciao
Phil