I feel that as nanopore has gotten better the shortcomings of flye have become more obvious.

I would've guessed Alice Roosevelt, but maybe I'm just telling on myself.

You forgot 'sudo'

I'm grateful every day for the opportunity to do fun and interesting work that hopefully helps the people of Alaska.

Yeah, I can see the loophole closures written into that book like battle scars.

I feel like this sort of hobby must have an encyclopedic rule book to prevent shenanigans.

I suspect that sequencing companies are primarily a means of tormenting bioinformaticians and the development and sale of sequencers is secondary to that goal.

This is really cool, looking forward to checking these out.

He sounds delightful.

Perhaps this makes me a bit of a luddite, but I am never going to upload my data to your website. I don't care how cool your new bioinformatic tool is.

Very nice! Looking forward to testing this in the near future.

Most of the samples were at 40-60x, but we had a few between 20-40x that still generated good assemblies. I was using 20x as a floor, but may need to reevaluate that with the current Q scores.

These were clinical isolates grown as pure cultures. RBK114 library, basecalled with sup v5.0, assemble with flye, map with bwa mem, polish with polypolish.

48 samples with single contigs between 1.8 and 2 Mbases at >20x coverage.

18 had 0 changes
15 had 1 change
7 had 2 changes
8 had 3-6 changes

Bioinformatics hot take: If you make a tool, I appreciate your work. If you make it in R, I also kinda hate you.

I just ran polypolish on about 50 bacterial genomes(~1.9 megabases each). The vast majority of those assemblies had 0-2 changes.

It's honestly hard to justify running the miseq at this point. Particularly since our wetlab strongly prefers the ONT library prep.

On more than one occasion I've thought of a cool idea for a bioinformatics program only to realize @torstenseemann.bsky.social already wrote it a decade ago.

Is there a bluesky version of "I'm in this post and I don't like it"?