joint interpretation of smudgeplot and GenomeScope
Every time you run smudgeplot, run GenomeScope as well. The two plots must make sense together. I will explain here a bit better why it's so important.
Smudgeplot requires genomic kmers to pair, especially for very homozygous genomes with many duplications smudgeplot can get confused. Furthermore, smudgeplot is very dependent on the cutoff L, if chosen too low, kmers with sequencing errors can be very detrimental to the analysis, conversely, if chosen too high, real genomic kmers might be missed which can also lead to wrong conclusions. GenomeScope does not always provide clear answers on "which peak is which" in the kmer spectra, which is the reason why one might be interested in running smudgeplot.
Smudgeplot gives one an idea, which peaks in the k-mer spectra have similar k-mers. In the following scheme you can see how the smudgeplot smudges directly relate the to respective k-mer spectrum.
Try it yourself, perhaps on