Closed JohnUrban closed 2 years ago
I think we should have a refresh of models in the coming days.
any timelines on when models for guppy 6.1.2 may be available.
bumping this to see if there is a timeline?
I think we should have a refresh of models in the coming days.
Any updates? Its over a month now.
Looking at Guppy changelogs and doing md5sums of the model files in Guppy 5.0.7 and 6.1.2 suggests that the R9.4.1 DNA models haven't changed, so the r941_min_sup_g507
model is correct for use with Guppy 6.1.2.
There are new models for R10.4.1 in medaka v1.6.1.
Ouch. That is brutal to find out considering the consensus sequence output by Flye is consistently outperforming further polishing with Medaka in my tests.
@JohnUrban is your test data and and metrics available to share? I don't actively work on medaka myself anymore but there are people who would like to examine your data.
Hello,
I have 5-10X coverage ultra-long reads (N50 120 kb) generated recently using R9 Spot-On flow cell with the MinION.
I basecalled on windows with Guppy GPU basecaller v6.1.2, and assembled with Flye v2.9-b1774 with 1 round of Flye polishing ( I am also trying 3 rounds). Likely owing to the length of the reads, I generated a decent assembly despite the low coverage (1-2 Mb NG50).
I then did further polishing with Medaka v1.6.0 (installed this week with Conda). However, there was no model that exactly matched my situation. The model I chose was
r941_min_sup_g507
. Will this be ok? Do you plan on releasing ar941_min_sup_g612
model? Or would there be little difference with the g507 model?Best,
John
p.s. available models were: