Many congratulations to Dr Kiarash Jamali @kjamali.bsky.social for successfully defending his PhD thesis today! 🥳 Also many thanks to Jonas Adler and Randy Read for examining him. I expect great things from Dr Jamali in his career, so do keep an eye on him. #ModelAngelo
Model building in high-resolution #cryoEM maps is more fun when it's a protein purified from its native source and nobody knows the correct sequence. 😅
#ModelAngelo to the rescue!
Reading supplementary information of papers sometimes gives nice surprises: I just noticed that two of my EMDB entries were used to train #BlushRegularization in #RELION 😮
So I could not resist the curiosity to check the list of EMDB entries used to train #ModelAngelo, but this one has none […]
Can you solve the structure and use #ModelAngelo?
It would be helpful to hear some more stories about how useful #ModelAngelo has been in your #cryoEM structure determination efforts. Has it changed how you build your models, perhaps even using other programs? 🤗
Would anyone in the blue hive perhaps have an idea what this could be? About 60 Angstroms across, probably made of 4x9 alpha-helices. Found in hexagonal arrays in mouse brain preps... (Would solve structure and use #ModelAngelo, but only have top views)
I can also just add to this. With @kjamali.bsky.social #ModelAngelo we were able to identify new proteins that are part of our complexes, where nobody had an idea they even existed. I complete changes the way how one can do research on non-standard organisms.
@kjamali.bsky.social I can’t tell you how much of a change #ModelAngelo has brought to our #cryoEM work. The time savings are just nuts. Complete novices to model building now get excellent starting points for refinement (in #ISOLDE) in minutes. It’s unreal. Thank you & @sjorsscheres.bsky.social
One more to follow on this side: Kiarash Jamali @kjamali.bsky.social is a PhD student in my group and the main developer of #ModelAngelo, which is a machine-learning tool for automated model building and the identification of proteins in #cryoEM maps.
Kiarash Jamali had initially shown that #ModelAngelo builds 77% of last year's (unique) EMDB submissions. We usually prefer to report impressive performance on its own over "my-software-is-x%-better-than-yours" comparisons. But the reviewer is always right! 🤗
We have updated the @biorxivpreprint.bsky.social of our #ModelAngelo paper. It now contains a more extensive comparison with other softwares. We build ~80% in maps with local resolutions <4A; the next best alternative needs <3A maps to do the same...
www.biorxiv.org/content/10.1...
Congratulations to Kiarash Jamali for winning the Max Perutz PhD student prize at the LMB for his development of #ModelAngelo for automated model building in #cryoEM maps! 🥰🥳
Too hard to resist the #ModelAngelo frenzy. Gave it a go using our small CDK-activating friend. Map computed from 4 hours worth of Glacios 2 data, live processing, 2 hours worth of RELION processing, and then automated modelling without sequence. The quality is quite impressive.
