Tristan Croll

LMI4Boltz: Optimizing VRAM utilization to predict large macromolecular complexes with consumer grade hardware AlphaFold2 has revolutionized structural biology by enabling the prediction of protein structures approaching experimental quality. AlphaFold3 extends this framework to support modelling broad biomole...

This is amazing work lead by Tom (@tlitfin.bsky.social) in my team. Always trying to get more for less and making computing efficient. So satisfying making science resources go further. Boltz2 is great for protein structure prediction and hopefully LMI4Boltz extends its use. Code on GitHub… 🧶🧬

Ha! Well… at least it got there in the end.

Sadly not… iirc there was one that achieved brief infamy a year or two back which had a negative overall Q-score.

Oh, my poor baby. What have they done with you??

Mashing Pumpkins

Oooh, I know lots of those!

Possibly the most evergreen Onion article in history: theonion.com/recession-pl...

“What is biology?”

A small, scared bat showing off its fangs.

This brings back memories… climbing at Mt Arapiles, near the top-out about 30m off the ground I came across a crack very like this, with an odd white spot deep inside. Got closer to look, and realised it was a teeny tiny pair of fangs… unforgettable moment.

I’m guessing your problem is the dependency on the “sanity” package. Unfortunately that stopped being maintained a while back, and won’t install in the latest Python version.

Several microscopic iron containing protein crystals in a variety of cookies (blue, yellow, orange, green)

Had a very long social media break after the last #SynchrotronMugMadness and the bird site's decent into horror.

So hello bluesky, I'm a staff structural biologist helping @imperiallifesci.bsky.social solve protein structures with #crystallography and #CryoEM.

Please enjoy these protein crystals 💎

🎶 When a muppet with teef finds a home on the reef, that’s a moray 🎵

(this is actually the old laptop I used for most of my original work on ISOLDE. Poor little thing… will try to look after it better in future)

A solid wall of accumulated dust. Somewhere underneath it is a laptop cooling fan.

Kids were complaining that their gaming laptop was getting a bit sluggish… 😳🤢

Interviewer: Can you explain this gap on your CV?

Evolutionary biology: Well, yeah… but then I’ll have two *new* gaps you’ll want me to explain.

Lady Cassandra O’Brien, last human alive in the year 5 billion (from Doctor Who).

Lady Cassandra??

Could of course fire up a “proper” docking package for that, but often this way seems quicker and easier (and helps get a better feel for the site, to boot). Take out the map, “isolde restrain distances” to maintain local conformation, then just play with it.

Both true. :) Although if it’s a novel ligand (or just something not in the CCD) the first point isn’t such a big issue. I don’t use it *that* often, but I do occasionally use it quickly assess hypotheses (“I wonder if{ligand} might plausibly fit this pocket?”).

Architecture of the 1.3 MDa complex of human UBR4 (2 x UBR4/KCMF1/CALM1). Picture shows top and side views of the cryo-EM density map and the modeled structure, colored by protein components and domains.

🎄 xmas preprint 🎄 we are excited to share our cryo-EM structure of UBR4 in complex with KCMF1 and CALM1. the PQC ligase forms a massive ubiquitination arena, primed to amplify ubiquitin chains (E4 activity) and boost degradation of defective proteins. www.biorxiv.org/content/10.1...

Still trying to work that out, I’m afraid. While we have some interesting “ISOLDE-adjacent” stuff in the works, my day job currently doesn’t leave a lot of (well… any, really) time free for development of ISOLDE itself. Exploring options to support it in a more sustainable way.

One nice quality-of-life addition: all “open” commands now have an optional “cofr” argument. So for example…

open ligand.mol2 center cofr

or

open {smiles string} from smiles center cofr

… will put the ligand on ISOLDE’s crosshairs ready to move around and merge in.

While I haven’t had a chance to add anything really substantial to ISOLDE, the @www.cgl.ucsf.edu team have added a bunch of great new stuff to ChimeraX itself.

For people who managed to install older versions of ISOLDE into ChimeraX and found they promptly failed… sorry, that’s my fault. I didn’t RTFM about Python’s ‘~=‘ specifier. Thought “~=1.8.rc0” meant “>=1.8.rc0”, ==1.8.* when it actually means “>=1.8.rc0, ==1.*”.

🙃

Anyway, update to fix.

You know what they say about never releasing on a Friday… what does it say about me that I’m releasing on the last Friday before Christmas? 😬

Anyway, ISOLDE builds for ChimeraX 1.9 are now up. Enjoy!

HPV Infections: Friend and Foe

An interesting finding that some types of HPV infection might be preventing cancer (as opposed to the ones that can cause it).

Also features G.K. Chesterton and a brief cameo by RFK Jr.

Man… the fact that there’s a *second* ring galaxy visible in the gap? The sci-fi story (where these are the still-growing explosions from the past firing of some unimaginably powerful galaxy-killing beam weapon) practically writes itself.

I’m afraid I honestly don’t remember that. 🤷‍♂️

I remember Jane Richardson and her team doing a fair bit of work trying to sort out “total junk” from “partially correct” down in the very-low-pLDDT regime. Don’t know if/where it was published, though.

Rare sighting of wild G-protein in Virginia