Hello, I am working with pacbio data derived from a benthic, filter-feeding marine invertebrate. Because of this, it was not possible to get a completely clean DNA sample from this species. The genome size is estimated to be ~250 Mb and I have roughly 200x coverage from sequencing data. The sequencing data itself has an average read length 8 kb and read n50 15 kb.
So far I've gotten the best assemblies with canu, although when I check for contamination using blobtools2 there is a low presence of contigs taxonomically assigned as bacteria as well several microscopic animals (none of which belong to the same phylum as the host). Because of this, I am worried that reads from these contaminating organisms may have been incorporated into the host contigs. I was wondering if there are any recommended tools to screen the draft assembly of my host organism for the presence of possible chimeric contigs? That way I may know if I need to attempt a new assembly with stricter overlap parameters.
I can only speak from a bacteria perspective, but 2 common tools for read binning are
krakenandcentrifuge. As long as you feed either of them an appropriate reference database, I think it should work.You could try assembling with a more strict parameter set. It really depends how close the organisms are as to how likely it is that they have shared kmers to create chimeric contigs.
If your marine invertebrate is very close to some known species you could try reference based assembly (but I guess it is complecated as eukaryotes have this huge genomes that are almost doomed to be incomplete at the first assembly). This worked for me. I would also recommend that you try several tools and compare the assemblies with mauve. For pacbio, Flye worked really better than canu in my case. You could also try abruijn that is not bad.