Lim Heo
@huhlim
Started a new position at Bristol-Myers Squibb!
CG models can be quickly (~s) and very accurately (<0.5 A, ~2 MolProbity score) reconstructed to atomistic structures using ML. "One particle per residue is sufficient to describe all-atom protein structures" biorxiv.org/content/10.110…
Some words of wisdom from this year's chemistry laureate David Baker, whose work has made it possible to design new proteins from scratch. Read more and join us on tomorrow 8 December to hear his Nobel Prize lecture: nobelprize.org/prizes/chemist…
Our team at Bristol Myers Squibb in Cambridge, MA is looking for a student for a Spring 2025 co-op. …stolmyerssquibb.wd5.myworkdayjobs.com/BMS/job/Cambri…
Generative Modeling of Molecular Dynamics Trajectories | NeurIPS 24' |📸"molecular video" - MDGEN, a flow-based model for MD trajectories, with various capabilities (forward simulation, interpolation, upsampling, and molecular design) - Tokenize structure as Roto-Translation and…
We have trained ESM3 and we're excited to introduce EvolutionaryScale. ESM3 is a generative language model for programming biology. In experiments, we found ESM3 can simulate 500M years of evolution to generate new fluorescent proteins. Read more: evolutionaryscale.ai/blog/esm3-rele…
Our team at BMS is seeking an expert in #NGS data analysis and #machinelearning for antibody therapeutics. If you're interested in this opportunity, don't miss out! #hiring linkedin.com/jobs/view/3954…
for the folks who missed @sokrypton's excellent presentation and discussion of #AlphaFold3 last week, you can now check out the recording 👇 youtu.be/qjFgthkKxcA
Introducing FoldFlow-2, a new SOTA sequence-conditioned protein generative model! work w/ @DreamFoldAI James V @FatrasKilian Eric TL @PabloLemosP @riashatislam @ChengHaoLiu1 @jarridrb @tara_aksa @mmbronstein @AlexanderTong7 @bose_joey Arxiv: arxiv.org/abs/2405.20313 1/8 🧵
Sharing an early preprint of my Microsoft AI4Science summer internship project. We developed SE(3) flow matching for protein backbone generation. Compared to SE(3) diffusion, we find our method achieves higher designability, faster sampling, with a way simpler implementation. 1/8
South Korean scientists’ outcry over planned R&D budget cuts. In a nation with historically high levels of spending on research and development, the proposed cuts have provoked a strong response. go.nature.com/3ZLIcw1
“SaProt: Protein Language Modeling with Structure-aware Vocabulary” An ESM1b-sized PLM pre-trained on structural data that outperforms other PLMs on various tasks biorxiv.org/content/10.110… github.com/westlake-repl/…
We are looking for a Project Manager 🧑💼👩💼to help us with the MDDB project implementation. Do you have knowledge about databases, data management, MD simulations and project management skills🔖? Would you like to join the team? Last days to apply 🗓️➡️ recruitment.irbbarcelona.org/jobs/2447575-i…
Very excited that one of my major postdoc works has been just published online @PRX_Life ! The title of the paper is: From Effective Interactions Extracted Using Hi-C Data to Chromosome Structures in Conventional and Inverted Nuclei. More details in the following thread! [1/n]
Hi-C, or genome-wide chromosome conformation capture, experiments report the interaction probabilities of arbitrary locations on genomes. Using Hi-C data, this #PRXLife study crafts a polymer parameter-free model that reproduces chromatin structures. 📝 : go.aps.org/3KJO7eU
The ESM models (including ESMFold!) have all been ported to @huggingface and will remain there even though the ESM team has been disbanded. We have example notebooks (look under 'Biological Sequences') if you've never tried it before! huggingface.co/docs/transform…
RIP ESMFold server 😢 (anyone has extra storage to archive all the predicted structures and pretrained models before @MetaAI pulls the plug)?
Statistical Mechanics: Theory and Molecular Simulation, 2nd Edition, now available on Amazon [amazon.com/Statistical-Me…] and will be available on Oxford University Press later this month for the electronic version, October for the hardcover version.
Did you know that #AlphaFold can be used to predict the relative populations of protein conformations AND their changes in response to stabilizing or destabilizing mutations? Check out our pre-print showing this unexpected but very useful feature: arxiv.org/abs/2307.14470
New Preprint with @Bulat41292666 : AlphaFold2 models of all 437 human catalytic typical protein kinases. Catalytic=not pseudokinases. "Active"=capable of binding ATP, Mg, and SUBSTRATE. Key step was structural bioinformatics of PDB structures of 40 substrate/kinase complexes.
AlphaFold2 models of the active form of all 437 catalytically-competent typical human kinase domains biorxiv.org/cgi/content/sh… #bioRxiv
"이게 그냥 국민들이 알아서 살아가고 있는 것 같은 느낌이 드는 거예요. 그러고 그냥 다 떠밀어요. 전에는 그래도 누군가 책임지고 사임을 하든가 하는데 지금은 너무 당당해요." 허망한 죽음들 그리고 아무도 책임지지 않는 풍토를 꼬집는 뉴스의 타이틀인 '일상된 추모'라는 말이 너무 먹먹하다.