Difference between reading concatenated PE reads into jellyfish and merging paired end .jf outputs independently

Question

[joshua.j.wright]

I have been finding my way through kmer counting some illumina paired end read sequence data for my PhD. When putting the data through jellyfish I have found massive difference between the outcomes after further downstream analysis i.e GenomeScope.

Attached is the images of the downstream analysis with concatenating a library (four files) before going through jellyfish and using jellyfish merge to merge the .jf outputs of each file independently .

The library consists of 288.4m 150bp reads all trimmed to Q25.

Is this merging the .jf outputs the best way to use the software?

Best wishes,