I have a fairly simple objective. Our group has sequenced 2 unknown organisms and 1 known organism (all eukaryotes). We suspect they may be closely related but we're not sure. My idea to check this was to use Kraken2 with a custom eukaryotic database and perform k-mer classification on all the reads. I was hoping to get an output that would show me how similar the organisms were, but I came across this article that showed that kraken2 struggles at the species level as the refseq database continues to grow (https://genomebiology.biomedcentral.com/articles/10.1186/s13059-018-1554-6). Because this is a lowest common ancestor program, I'm worried it will just output the same LCA at the genus level for all 3 organisms, in which this approach isn't suitable to figure out the similarity between all the organisms.
I'm wondering if anyone has any better suggestions for how to go about this study?
You could compare the three organisms directly to see if they appear to be similar by using a kmer based tool like sourmash.
Such an analysis may give incorrect results if contamination is present.
In event of contamination any/all analyses are going to have problems.
GetOrganelle, which I recommended, will make the consensus sequence of the mitochondrial genome and nuclear rDNA. If the contamination is not extremely abundant, it should not pose a problem.