themaklin.bsky.social
I do statistics, bacterial (meta)genomics and bioinformatics software development.
Researcher @ University of Oslo
99 posts
1,358 followers
405 following
Prolific Poster
comment in response to
post
How your ChatGPT-generated email finds me
comment in response to
post
The publisher has cut their costs by outsourcing to this company, the company has cut their costs by using AI/low-paid staff instead of paying for a proper job, while I’ve spent hours & hours fixing the manuscript, so all the extra labour from cost-costing has fallen on me, the unremunerated author
comment in response to
post
Thank cmake for that
comment in response to
post
That's a very interesting project, thanks for pointing it out!
comment in response to
post
Oh nice, thanks, I'll have to give this a try next time I work on the browser UI
comment in response to
post
It's surprisingly fast and very usable for the scale these typically operate at! :) we didn't optimize sbwt at all for this either
Biggest current issue is I couldn't figure out how to make a "download" of the results (can't access the filesystem) so they have to go through the clipboard.
comment in response to
post
All our algorithms run in the browser, so try our tool at maklin.fi/kbo/ There is no server-side processing involved, so the data never leaves your computer.
More theory about the algorithms in the preprint. Thanks @jnalanko.bsky.social, Elena Biagi, Simon Puglisi for working with me on this!
comment in response to
post
The implementation of kbo comes in three different flavours:
- core Rust library, kbo
- command-line client, kbo-cli
- in-browser client, kbo-gui
These are available from GitHub github.com/tmaklin/kbo. The command-line client can be installed from bioconda or from crates.io using cargo.
comment in response to
post
We also improved the precision of SNP and INDEL calling compared to other tools by using a specialized variant calling algorithm based on the suffix match lengths of both query vs. reference and reference vs. query.
In reference-based mapping, we are within 1-2% of using sequencing reads