Xin@QuantumChem
@xinquantum.bsky.social
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
#compchem #compchemsky
November 2, 2025 at 7:49 PM
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
#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
#compchem #compchemsky
November 2, 2025 at 1:40 AM
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
#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
#CompChem #ChemChemSky
October 22, 2025 at 7:18 PM
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
#CompChem #ChemChemSky
I'd like to name it as "QM/MM".
August 13, 2025 at 8:14 PM
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*.
The newly released open-source model from OpenAI can be tried on TractoAI playground for *FREE*.
TractoAI | Cloud platform for AI and Data teams
Run AI and Data workloads in the cloud at scale
playground.tracto.ai
August 6, 2025 at 8:25 PM
playground.tracto.ai/playground/n...
The newly released open-source model from OpenAI can be tried on TractoAI playground for *FREE*.
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.🪜
#MachineLearning (ML) for orbital-free density functional theory (OF-DFT) and OF-DFT for ML. Manzhos and coworkers tackle the kinetic energy functional in this new #MachineLearningScienceandTechnology article. #compchem #CompChemSky Read it here: bit.ly/45eGM0f
August 1, 2025 at 10:03 PM
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.
July 27, 2025 at 8:04 PM
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
#CompChemSky #CompChem #GPU
July 23, 2025 at 8:58 PM
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
#CompChemSky #CompChem #GPU
Kvass and Khinkali
July 14, 2025 at 12:38 AM
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
Inspired by www.youtube.com/watch?v=mDyt...
#compchem #compchemsky
July 12, 2025 at 6:06 PM
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
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...
#compchemsky #compchem
colab.research.google.com/drive/1g1PiW...
Google Colab
colab.research.google.com
July 9, 2025 at 6:23 AM
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...
#compchemsky #compchem
colab.research.google.com/drive/1g1PiW...
This is a very interesting package. I will try this on colab notebook.
Wrote a bit of code to quickly analyse vibrational modes from QM calculations! #compchem
github.com/aligfellow/v...
Returns a list of internal coordinates associated with the vibrational mode, can be used from command line on xyz trjfiles or orca or gaussian outputs etc.
github.com/aligfellow/v...
Returns a list of internal coordinates associated with the vibrational mode, can be used from command line on xyz trjfiles or orca or gaussian outputs etc.
GitHub - aligfellow/vib_analysis: Quick analysis of xyz vibrational mode trajectory
Quick analysis of xyz vibrational mode trajectory. Contribute to aligfellow/vib_analysis development by creating an account on GitHub.
github.com
July 5, 2025 at 3:00 AM
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
#compchemsky #chemsky #compchem
July 4, 2025 at 12:14 AM
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
#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/...
pubs.acs.org/doi/10.1021/...
June 27, 2025 at 3:28 AM
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/...
pubs.acs.org/doi/10.1021/...
June 26, 2025 at 4:30 AM
Reposted by Xin@QuantumChem
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
Join us in Zoom: us06web.zoom.us/j/8607128794...
#compchem #dft
Join our Cloud HD Video Meeting
Zoom is the leader in modern enterprise cloud communications.
us06web.zoom.us
June 14, 2025 at 12:44 PM
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
Join us in Zoom: us06web.zoom.us/j/8607128794...
#compchem #dft
Reposted by Xin@QuantumChem
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
#compchem #CompChemSky
June 11, 2025 at 10:53 PM
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
#compchem #CompChemSky
An interesting opportunity to help the horseshoe crab (more closely related to spiders). gothamist.com/news/new-yor...
New York beaches are dating hot spots for horseshoe crabs
Beneath a full moon in Jamaica Bay, horseshoe crab love was in the air.
gothamist.com
June 12, 2025 at 1:16 AM
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
#compchemsky #chemsky
June 4, 2025 at 3:52 PM
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
#compchemsky #chemsky
JupyterNotebook: github.com/myzzzz6/AI4C...
May 22, 2025 at 2:45 PM
JupyterNotebook: github.com/myzzzz6/AI4C...
Get Started with Meta’s Molecular AI Model UMA on Google Colab
open.substack.com/pub/quantabr...
open.substack.com/pub/quantabr...
Get Started with Meta’s Molecular AI Model UMA on Google Colab
Meta and collaborators released the molecular foundation model UMA (Universal Model for Atoms) and its training dataset OMol25 (at ωB97M-V/def2-TZVPD level) this month.
open.substack.com
May 22, 2025 at 2:41 PM
Get Started with Meta’s Molecular AI Model UMA on Google Colab
open.substack.com/pub/quantabr...
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
As a person without CADD knowledge, I participated in the @polarishub.io ligand pose challenge. And use ChemOrchestra platform to achieve a 17.39% hit rate (RMSD < 2Å) with an average RMSD of 4.94Å. #CompChemSky #DrugDiscovery#VirtualScreening#ComputationalChemistry #MedChem #CADD #AIDD
March 21, 2025 at 9:23 PM
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.
Godot 4.4 launched last week to a great reception!
We’re now preparing 4.4.1 to address regressions and ensure a smooth upgrade for all users. Help us test the release candidate and report any issues you encounter.
godotengine.org/article/rele...
We’re now preparing 4.4.1 to address regressions and ensure a smooth upgrade for all users. Help us test the release candidate and report any issues you encounter.
godotengine.org/article/rele...
Release candidate: Godot 4.4.1 RC 1 – Godot Engine
Godot 4.4 was released 10 days ago, and as is customary with such major feature updates, we have a late harvest of regression fixes to offer!
godotengine.org
March 15, 2025 at 6:00 AM
It's suitable for building AI agents.