Software and services offered at Nostrum revolve around one crucial element: our infrastructure. We offer an accessible computing environment that follows HPC (High Performance Computing) standards to bring efficiency and affordable biomolecular computation to everyone. This infrastructure is ready to run the main molecular modelling and artificial intelligence software and also supports big computational workloads that require high scalability. In this last scenario, Nostrum has access to HPC environments on the cloud, where software is carefully compiled to run in the most effective way, to extend its resources to thousands of CPU cores and tens of GPU cards. We can process more than 50K CPU core hours every day.
The NBD Suite is the product developed by Nostrum that aims to offer solutions on the following fields:
Technology: the combination of different technologies to offer, among others, a high performance computing infrastructure, high cybersecurity standards, and a simple and pleasant user experience.
Molecular modelling: the union of our proprietary software, open source tools and third-party licensed programs into workflows of modular blocks to be applied to drug discovery and enzyme engineering studies.
Artificial intelligence: a collection of state of the art techniques to manage the data used and/or obtained in computational models to predict, generate, and explain it beyond the limits of molecular modelling.
NBD Suite Web App
The NBD Suite Web App offers free access to a simple web interface where a limited number of applications and resources of our infrastructure is provided. This service allows users to submit different types of biomolecular simulations and display results. Users can run quick induced fit docking simulations combining rDock with PELE to identify the best binding modes of a ligand in a protein cavity. Artificial intelligence models are also available to predict ADMET properties on sets of small molecules. These applications are available through our corporate website for academic users and software evaluation purposes.
Our Protein Energy Landscape Exploration (PELE, Borrelly et al. 2005 J. Chem. Theo. Comput. 1(6), 1304-1311) tool combines protein structure prediction algorithms and Metropolis Monte Carlo techniques to efficiently tackle tasks like predicting binding sites, docking pose refinement and ranking, protein surface exploration to look for binding sites or modelling exit path of a ligand. One of the main characteristics of PELE is its ability to explore the degrees of freedom of both ligands and proteins while keeping a low computational cost.
pyDock is a protein-protein docking algorithm (Cheng et al. 2007 Proteins 68, 503-515) that uses electrostatics, desolvation energy and a limited van der Waals contribution to score rigid-body docking poses (pyDockSER). pyDock participates in CAPRI (Critical Assessment of PRediction of Interactions) under the name of “Fernandez-Recio”. You can evaluate its performance among the rest of participants visiting the summary provided by Yang Shen from Vadja Team.
In order to perform most of rational small molecule design tasks Nostrum is endowed, among a series of different software packages, with one of the most highly regarded platforms in the industry, the Schrödinger suite of programs. Moreover, Nostrum counts with other non-proprietary software programs where we have extensive expertise such as AMBER20, GROMACS, AutoDock, rDock or Chemical Checker to complete its portfolio of technologies.