Atomics AI

[Available On-Demand]
Search General Info
Search Education
Search Partnering Companies
We use quantum mechanics and AI to predict protein-ligand binding, select better drug candidates, and accelerate drug development. Our AI learns from quantum mechanical (QM) simulations of molecules which is 4x more accurate than conventional approaches on the market. Also our AI predicts bindings 1,000x faster than QM. More specifically, in order to assess the interactions between molecules, we compare the binding energies of molecular duplexes in S22 data set. The S22 set includes hydrogen-bonded type, van der Waals type and mixed type molecular bindings. The energy resolution of our approach is about 4 times better than that of conventional approaches on the market. In addition, AI runs 1,000x faster than QM and predicts the energies accurately. The mean absolute error is less than 0.07 kcal/mol. We have proved the accuracy & efficiency of our QM+AI approach for small organic molecular duplexes and now work on amino acids to predict protein-peptide binding.
Company Type:
Company Website:
Company HQ State:
Not Provided
Company HQ Country:
Korea - South Korea
Year Founded:
Main Therapeutic Focus:
Lead Product in Development:
Innovative drug discovery tech leveraging QM+AI
Development Phase of Primary Product:
Number Of Unlicensed Products (For Which You Are Seeking Partners):
Not Provided
Atomics AI