Our Platform
The Alivexis’ Drug Discovery Platform integrates and combines our expertise in cutting-edge supercomputing technologies, advanced structural analysis methods, synergistic technology / drug discovery collaborations with our partners, and a globally agile networked operational model. The result is a radically efficient, modular, and scalable drug discovery engine.
ModBindTM Simulation Engine
Our ModBindTM* computational platform predicts ligand efficacy at high accuracy and speeds enabling our drug discovery programs to advance rapidly. ModBindTM uses Molecular Dynamics (MD) simulations and a novel and theoretically rigorous statistical mechanics theory to predict absolute free energy of binding and off-rates of drug candidate compounds. Our method is highly efficient, allowing for predictions up to 1000X faster than current state of the art free energy methods. When ModBindTM is used at scale on GPU supercomputers, we have the ability to predict 10,000+ ligands in days or less. This allows us to use the method in early hit finding for novel targets where we have demonstrated 10X the number of active chemotypes compared to conventional virtual screening, as well acceleration of hit to lead and lead optimization. We only make the compounds that have a high likelihood of being potent by screening them with ModBindTM, accelerating the discovery of new medicines. While ModBindTM is built on physics-based calculations, it also shows great promise to significantly advance the utility and accuracy of AI in drug discovery. AI models for biological target affinity have long suffered from a lack of high quality data, due to the cost and time it takes to generate experimental data. For this reason, we believe the true promise of AI in drug discovery has not been fulfilled. ModBindTM can be viewed as an in silico assay for ligand activity which can rapidly and consistently generate orders of magnitude more data than experimental data accumulated over the years of a typical drug discovery project. We believe training on ModBindTM generated in silico data will provide a great leap forward in the accuracy and applicability of AI models to predict drug-target affinity.
* https://doi.org/10.1021/acs.jcim.4c01805
Advanced Structural Analysis
Alivexis invests significantly in obtaining unique, exclusive, and high-quality 3-dimensional drug target structures, allowing us to fully leverage our supercomputing technology platform and providing our portfolio with a unique advantage.
In addition to active efforts in pursuing target structures through protein crystallization, we are working closely with top-tier collaborators with state-of-the-art Cryo-EM capabilities, enabling us to determine structures for targets that were inaccessible through traditional crystallization approaches. We perform modeling of structures using various technologies, including AI.
We already possess over 100 unique drug discovery target structures and binding structures, including entirely novel entities, and we are expanding our repertoire of drug target structure information on a daily basis.
Synergistic and Strategic Collaborations
In close collaboration with our partners, we are actively building innovative combinations of our platform with various other experimental platforms, techniques, novel active lead molecules, and biological approaches.
Our collaboration with PeptiDream, Inc., is a unique example in which we combine PeptiDream’s Peptide Discovery Platform System (PDPS) with Alivexis’ platform to identify, analyze, and pursue novel approaches towards targets and compound profiles that have previously been considered difficult. Under this collaboration, by using selective and potent peptides obtained through PDPS as probes, we are discovering new target structures and compound binding sites, aiding in the search for new small molecule compounds for various targets at Alivexis. As a result, we have already identified clinical candidate compounds for several of our lead projects.
In addition to our PeptiDream collaboration, we are working closely with multiple partner companies and academic institutions at the forefront of various research and technology areas to pursue novel approaches to small-molecule drug discovery. Among those, the collaboration with Fujitsu leveraging their advanced technology in high performance computing has succeeded in further accelerating large-scale virtual screening of more than 10 million compounds on “Fugaku”, one of the world fastest supercomputers, which was jointly developed by RIKEN and Fujitsu. In addition, we have been continuing collaboration with top academic scientists, for example in Yamanashi University and Hokkaido University both in Japan, to explore the novel therapeutic opportunities in the diseases which still have significant unmet medical needs such as allergic disorders and vasculitis. Some of our achievements in the collaborations using our proprietary compounds have been already published in top-tier scientific journals including Nature Communications, validating our science in drug discovery.
Globally Networked Operations
Alivexis is building an agile and scalable drug discovery research operation system through a lean internal team of Alivexis’ experienced scientific leaders, working closely with multiple industry partners and prominent scientific advisors in specialized disease biology areas.
We capitalize on our rich professional networks built over many years in the pharma industry and our global experience working with partners to access specialized functions, experts, and companies at the cutting-edge of biology and technology to advance Alivexis’ drug discovery pipeline.
The result is an efficient networked operational model with the ability to rapidly and flexibly mobilize, and pivot and adapt critical R&D functions as our discovery programs evolve.