I have problems interpreting the results from our assemblies. We have sequenced an animal with nanopore and Illumina reads. I first performed Canu v1.7.1 assemblies of subsets of the nanopore reads to see effects on the outcome. Here I generated individual assemblies with all reads >1kb, >2kb, etc. up to 20kb (with a jump from 15kb to 20kb straight). Afterwards, I corrected each Canu assembly with the Illumina reads each with 10 iterations of Pilon error correction (using Bowtie2 mapping). Then I used Bedtools to get histograms for each mapping from the bam files. The results of the per-base coverage plot can be seen on the attached graph (cov on the x- and millions of bases on the y-axis; the number in brackets behind the sets is the rough nanopore coverage based on the 1kb assembly length).
In the plot one can see coverage ratios of 0.25, 0.5 and 1. First I was interpreting the graph as a triploid organism, but this does not fit to the ratio. One would expect 0.33-0.66-1 for a triploid.
I also performed a mapping of the corrected & trimmed >1kb nanopore reads (generated by Canu) to the raw >1kb read assembly with bwa and also plotted a per base coverage.
Any Idea what is going on here? Any help in interpreting the results would be highly appreciated. This drives me nuts already...
Thanks for your help.