Question about error rate for correction

[mylena-s]

Hello! thank you for developing NextDenovo, it is working perfectly for me.
I am writing to ask you a question about the parameters for the run.

I am working with ONT reads, mostly obtained using the V14 chemistry of the R10.4.1 flowcells. The median quality of the reads is 20, and there is a high proportion of ultra long reads with lengths above 100Kbp, up to 800Kbp.

I also have other datasets obtained with the previous ONT flowcells that have a higher error rate. When I compare the results of the assemblies obtained with both technologies (but similar coverages), I find higher N50s for the older flowcells (12Mb vs. 6 Mb).

I know this could be coincidental, but I was wondering if there is any parameter that I could modify to account for the lower error rate or larger reads, and therefore improve assembly?

Thank you in advance!

Mylena

[Hu Jiang]

You could try increasing the value of -k -w in minimap2_options_raw and minimap2_options_cns, I'm not sure if that will improve the assembly, but it will definitely introduce the running time.