Andrew Rambaut (🐏🎀)

Coincidentally, the newly minted icon...

Submit an issue at github.com/artic-networ... and provide an example and some background and I will look into it.

I guess you mean you won't be using any programs where the developer tells you they used AI?

Post one as an issue and Claude and I will see what we can do: github.com/artic-networ...

Damn. Will rename it BearTree now.

Transit data? Where did it get that from?

I haven't even had time to design an icon.

For this sort of app – where bugs are apparent because you can see them and it is low-stakes because you are creating visualizations of scientific data rather than doing science – LLMs seem an extremely effective tool. PearTree was created in two weeks and I was doing other stuff at the same time.

This experience has been in interesting one - I think it helps that I know JavaScript quite well and feel I could have written this if I had had the time (with endless trips to StackOverflow). The code base is now large but generally easy to follow and well annotated.

Finally, I would like to be able to say this is the product of years of work but it certainly wasn't. I designed it, specified exactly how it should work, how the code should be structured and the frameworks to use but most of the actual code was generated by Claude Sonnet 4.6.

artic-network/peartree Contribute to artic-network/peartree development by creating an account on GitHub.

But happy to have requests and bug reports through GitHub issues github.com/artic-networ... and it is all open source code etc. so PRs also welcome. Including new themes and palettes.

artic-network/peartree Contribute to artic-network/peartree development by creating an account on GitHub.

There are features from FigTree that are missing some of which I plan to implement, some I am not (I am looking at you circular trees).

It can also handle unrooted trees such as those produced by the lovely IQ-TREE (iqtree.github.io) including bootstrap values. And you can do various re-rooting operations on these 'unrooted' trees.

For obvious reasons it is quite orientated towards BEAST trees (i.e., MCC or HIPSTR trees with posterior summary information about the nodes). You can calibrate a timescale, draw node bars etc. Colour things based on annotations etc.

These should work like normal desktop apps (whatever is normal to your OS, at least) but I haven’t tested the Windows or Linux versions because I haven't.

Releases · artic-network/peartree Contribute to artic-network/peartree development by creating an account on GitHub.

Whilst web apps are quite handy as they don’t need installing, they can be annoying (e.g., when you accidentally close the window). So you can also download desktop apps here: github.com/artic-networ...

There is a tutorial here if you want to take your time: artic-network.github.io/peartree/doc...

So I pleased to announce the conceptual spawn of FigTree: PearTree (acronym still to be finalised). If you want to dive right in it is hosted as a web app here: artic-network.github.io/peartree (click the “Example...” button for immediate candy and then click every button you can find).

I won’t be able (don't want) to implement any new features including ones I really wanted for my own work. Instead of trying to find time to fix/enhance FigTree I decided a more fun procrastination project would be to start it all again from scratch.

So I have been neglecting it and there are some functionality issues (usually to do with specific combinations of OS version and Java version). I still hope to find time to fix some of the most egregious outstanding issues but....

What I had always tried to do with FigTree is make it look and behave like a native desktop programme in whatever system it was being used in. This has become harder over the years because the UI components in Java haven’t kept track with how Mac OS and Windows looks and works these days.

rambaut/figtree Automatically exported from code.google.com/p/figtree - rambaut/figtree

The first version of FigTree was released nearly 20 years ago and it is still widely used (including by me). But there are currently 85 issues on the GitHub repo (github.com/rambaut/figt...) and some of them I don’t really like the look of.

In 2025, GISAID cut off access to many critical SARS-CoV-2 resources, incl. our outbreak.info.

A month ago, GISAID leveled accusations against us, Cov-Spectrum, Nextstrain, and members of COG-UK, which are false and misrepresent known facts.

Our joint response:

blog.outbreak.info/joint-respon...

Something very different and cool for my lab. Sequencing a 4.4mb TB genome using the Artic-style PCR amplicon method with >5128 primers in two reactions 🤯

Enables culture-free WGS for TB!

journals.asm.org/doi/10.1128/...

Anyone who missed their childhood vaccination (probably in their 40s to 50s) can get MMR on the NHS for free and it seems an increasingly a good idea to do so as incidence rises. You really really don't want to get measles as an adult.

That was like me - didn't get the vaccine as a kid (because it wasn't routine) but had no memory of getting the disease. I got vaccinated when there was an outbreak in the university - exactly the period when kids born during the Wakefield/Daily Mail anti-MMR campaigns were coming to university.

There is also the 'not yet vaccinated' (under 1y in the UK). Plus older people who may not have been vaccinated as a kid and haven't been infected because people around them have been vaccinated. Measles is so transmissible it can get all the hard-to-reach susceptibles.

Recapitulating the host species phylogeny seems like a different thing - suggests host specialisation of the microbes. Seems implausible to be due to vertical transmission which is what the 'microbiome' concept seems to be suggesting. But caveat again: not an expert.

But presumably people have found that in the past to come up with the concept of microbiome (not that familiar with the literature on microbial gut fauna to be honest).

Nice, stable, nutrient-rich environment with easy dispersal to new homes (lots of things eat poo - individual of the same species probably the least likely). If the host relatedness mirrored the microbial relatedness then that would be the more surprising finding in my opinion.