Our technology
Accelerating MD simulations using evolution and AI
Our mission
By understanding the molecular details and obtaining functional readouts from simulations we can computationally optimize drug leads towards a target
In effect, we can drastically reduce the amount of time and resources needed from identifying a target to preclinical studies.
Thanks to our dedicated team and computational resources we can deliver projects in fast timescales (2 – 12 months depending on scope and complexity)
Our mission is to revolutionize drug discovery, and bring rationality and molecular detail to our clients.
Technology rooted in 2 nobel prizes...
... and 20 years of research and development. Specifically, we combine the fields of computational chemistry and molecular simulations (awarded a Nobel Prize in 2013 for the development of multiscale models for complex chemical systems) with advanced modelling of complex physical systems (awarded a Nobel Prize in 2021 for the discovery of the interplay of disorder and fluctuations in physical systems from atomic to planetary scales). Add to these groundbreaking discoveries the deep expertise present in our team, were we have over the course of 20 years applied and developed molecular simulations for next-generation rational drug discovery.
Accelerating classical Molecular Dynamics simulations
Our solution is to employed advanced enhanced sampling algorithms that we are continuously improving - both to harness the power of on-the-fly neural network evaluation as molecular descriptors and effectivise the sampling algorithm by adaptively fine-tuning how new microstates are accessed.
Dimensionality reduction using information-greedy algorithms
As exploring alternative conformations of proteins or protein-ligand complexes, one needs to address the issue of dimensionality. Namely, proteins with N particles contain approximately 3N degrees of freedom, but not all fluctuations are interesting to explore. Instead, we rely on iterative exploration algorithms that are seeded by coevolutionary models.
By using this approach, we have been able to show that our technology is able to accurately predict functionally relevant states, even those that are only visible in presence of a subset of ligands (see Mitrovic et al 2023, J Phys Chem, Mitrovic et al 2023 elife or Yee, Mitrovic and McDonald et al 2023, BioRXiv). Through our methodology, we can therefore:
-Explore the conformational change linked to function
-Investigate how function is altered by drug-like molecules and mutations
-Use the detailed rationalization to design drugs or further understanding of the effect of disease mutations
Internal Development
We are dedicated to improving our methodology and bringing the next-generation of AI-powered drug discovery. Our novel discoveries are soon to be published as preprints.