Magnus Manske

You expect some honest, well-balanced information about myself? On my own page? Seriously? Better check out Wikipedia instead.

Now

Working as a software developer for the TIB, University of Hannover, since January 2025.

Then

Scientific software engineering at the Wellcome Sanger Institute, since April 2007. Malaria, then Parasites and Microbes.

Does

Turns large, awkward genomic datasets into software that biologists can use without noticing there is software involved.

Writes

Rust, PHP, JavaScript, C++, SQL, and a lot more. In roughly that order of current enthusiasm.

Known for

The first version of MediaWiki. GENtle. An unreasonable number of Wikipedia and Wikidata data tools.

Training

PhD in biochemistry, University of Cologne, 2006, magna cum laude. The thesis was a piece of software, which should have been a warning sign.

I am a biochemist working in bioinformatics at the Wellcome Sanger Institute in Cambridge. I joined in 2007 to work on malaria genomics, and moved to the Parasites and Microbes programme in 2019, where I build and lead work on the interfaces that let collaborators actually get at the data.

I write fast code, web code, and the occasional scientific publication. The full list lives on ORCID.

Roughly twenty years of shipping tools to scientists who did not ask for tools, they asked for answers. It teaches you things about requirements gathering.

As my PhD thesis I wrote GENtle, a free multi-purpose molecular biology tool. Despite wide usage, including NASA, and despite being open source, I eventually ran out of the one resource nobody grants you: time.

Thanks to the good folks at Synbiota it was reborn as a pure web application, with the code on GitHub.

And now there is a third one, in Rust: gentle_rs, under active development. It is a DNA and cloning workbench for both interactive use and automation, built around a single deterministic engine shared by the GUI, the CLI and the scripting interfaces. It plans and reviews Gibson assemblies, runs PCR, mutagenesis, primer-pair and qPCR design, imports GenBank, EMBL and SnapGene records, and renders protocol cartoons, lineage graphs, dotplots and gels straight from project state, so the figures are generated rather than drawn by hand.

Three implementations of the same idea across twenty years. Call it persistence, call it a diagnosis.

I have been involved with Wikipedia as both article and software author since before there was Wikipedia. I wrote one of the first versions of the software that became MediaWiki, plus the Cite extension that put references in angle brackets where they belong. They even named a day after me, which I mention now to get it over with.

These days it is mostly infrastructure: QuickStatements, Mix’n’match and its several thousand catalogues, PetScan, Reasonator, and a long tail of other tools that quietly move a large fraction of the edits on Wikidata. There is a blog about them.

Mix’n’match and QuickStatements both picked up Coolest Tool awards in 2019, WikiShootMe in 2021, PetScan in 2022, and MediaWiki received a USENIX STUG award in 2010. I am told this counts as external validation.

Talk to me about any of it on en.wp, de.wp, Wikidata, or Commons.

• GitHub the code, warts and all
• Codeberg more code, just what you need
• ORCID all my publications
• LinkedIn hire me and pay me more money
• Mastodon @magnusmanske@wikis.world
• ResearchGate something with science people
• Flickr look, pretty pictures

And probably some more, which I forgot. No Farcebook, though. I hate Farcebook.

If you cannot figure out how to contact me from all the links on this page, chances are you have nothing to say that would interest me. Darwinian spam filter!