Centuries of human progress led us to uncover The Double Helix. In the end, we realized that all living things share the same codebook. #science #biology

Yesterday I posted an incorrect comparison. I accidentally copied the TS3 structure into TS1 for UMA and Orb-V3, so the MLIP results for TS1 were wrong. Really sorry about that misleading! In the next post I will share IRC of DFT calculation and wB97M-V/def2-TZVPD results. #compchem #compchemsky

A failure case for deep-learning potential: the TS1 geometry is so weird that UMA and Orb-v3 underestimate the barrier by ~15 kcal/mol, likely because it far from the training set. Handling chemical reactions is extremely hard for general deep learning models. #compchem #compchemsky

UMA (small model) is very close to DFT M06-2X. Maybe I can deploy nanoreactors using UMA model for coarse reaction space sampling. This is a reactive trajectory of a compound with 24 Carbon atoms. #CompChem #ChemChemSky

I'd like to name it as "QM/MM".

playground.tracto.ai/playground/n... The newly released open-source model from OpenAI can be tried on TractoAI playground for *FREE*.

This old theory is revitalized by deep learning. Building accurate kinetic energy functionals (< 0.5 kcal/mol) may be harder than climbing to the top of exchange-correlation Jacob's ladder.🪜

Cunning ChatGPT bypassing 3D depth analysis, and calling the rdkit function to generate the coordinate of an Ibuprofen. Grok3 just failed this test and misrecognized it as glucose, as expected.

Boltz2 can be run on a $430 RTX5060 Ti (16G) card, although it's much slower than ~$30,000 H100 (RTX 5060 Ti: 1min 47s, H100: 18s), but it doesn't matter, because the cold-start time is about 1 minute for both cards. #CompChemSky #CompChem #GPU

Kvass and Khinkali

This is creepy, but lots of fun. Gemini Veo has a perception of spatial depth and can rotate the molecules. Inspired by www.youtube.com/watch?v=mDyt... #compchem #compchemsky

I ran a benchmark on a protein with ~3450 atoms. AIMnet2 takes 4.6 seconds, and UMA(small) takes 7.9 seconds. CUDA 12.4 on a Tesla T4. Additionally, the AIMnet2 checkpoint file is only 11.1 MB. The testing notebook link is shared. #compchemsky #compchem colab.research.google.com/drive/1g1PiW...

This is a very interesting package. I will try this on colab notebook.

UMA (uma-sm) combined with NEB successfully locates the transition state and path for a Ni-catalyzed reaction. UMA prediction lies between two DFT functional, and the transition state closely matches the result from a literature: www.nature.com/articles/nat... #compchemsky #chemsky #compchem

The relationship between Ki and IC₅₀ is described by the Cheng–Prusoff equation. And William L. Jorgensen et al. did an effective calculation for predicting potency in his paper during pandemic. pubs.acs.org/doi/10.1021/...

Guess the name？ #Cocktail #Liquor #Alcohol #cursor

2nd virtual lecture in the DFT&Application series starts today at 1500hrs CET (in 15 minutes!). Topics are Hartree Fock, SCF, and their connection to DFT. Join us in Zoom: us06web.zoom.us/j/8607128794... #compchem #dft

Hey Jan, I just completed UMA benchmark tests on large molecules (biomolecules, macrocycles, and more): quantabricks.substack.com/p/uma-testin... In short, when atoms number below 700, uma is more expensive. Otherwise, GFN2-xTB is more expensive #compchem #CompChemSky

An interesting opportunity to help the horseshoe crab (more closely related to spiders). gothamist.com/news/new-yor...

Surprisingly, the UMA (by meta) model captures the bond-breaking process in the carbon nanotube stretching test, similar to GFN2-xTB. I hadn't expected deep learning models to extrapolate over bond lengths. #compchemsky #chemsky

It's a great time to participate in the horseshoe crab protection activity (helped with counting and tagging). I had only heard stories about the blue blood legend in evolution before, but this time I saw it with my own eyes. #aquatic #science #NYC #drug

JupyterNotebook: github.com/myzzzz6/AI4C...

Get Started with Meta’s Molecular AI Model UMA on Google Colab open.substack.com/pub/quantabr...

Interesting results for physical model, average RMSD<5A meaning that the ligands are plugged into the correct docking pocket. Unlike the AI methods, the charge models, protein ionization state or hydrated docking should be considered. This competition is also hosted by @asapdiscovery.bsky.social

It's suitable for building AI agents.

go.bsky.app/52czHRA #StarterPack for #CADD (computational-aided drug design) enthusiasts, open to researchers, students from both industry and academia. #compchemsky #compchem #chemistry #drugdesign

LLM-based ESM3 has just been released. Protein structure predicting is just a starting point of workflows. More functions should be implemented.

Seeking Image-to-SMILES conversion for local tiny device deployment. Deep learning offers accuracy but is too heavy. Are there lightweight alternatives without deep learning? Mathpix (SaaS) appears best, but closed-source and need internet.

Stretching of a carbon nanotube, a DFTB simulation. Do you know why I placed a nitrogen atom at the center? #CompChem #CompChemSky #Science #Chemistry

Lower speed collision of C60, like two elastic balls. (DFTB+) #CompChemSky #CompChem #Chemistry #Physics @filtei.bsky.social

It looks like a Christmas Wreath

Collision Between Two C₆₀ Molecules: A Violent Encounter (An xTB-GFN2 Simulation) #CompChemSky #CompChem #Chemistry #Physics

Conchiglie with Kale and ham, The center part of the shell is raw and hard, but tastes good. Is it authentic? #Italian #Italianfood

Simulation of a violent collision between two molecules. How many reactions can you find? Simulation using GFN2-xTB. On a preowned desktop computer. It takes about 20 minutes. #compchem #Chemistry #computing