Project description
Innovative machine-learning platform for prediction of potential candidates in drug discovery process
The early stage in the drug discovery process has the greatest impact on the success rate of new drug candidates. Computer-aided drug design (CADD) is the most cost-effective tool for drug discovery that allows the screening of large databases of molecules and simulates their interaction in vivo. However, simplifications in the simulations to achieve acceptable computational time often result in low success rates in finding the bioactive molecules. Spain-based company Pharmacelera aims to introduce a new computing standard for CADD using a machine-learning platform for the accurate prediction of molecular properties of candidate molecules and relevant molecular interactions. The EU-funded MolPredict project will be conducting market research and technical exploration to assess the feasibility of the endeavour.
Objective
The competitiveness of the R&D process in this industry has been steadily declining during the past two decades. Whereas in 2000 the average cost for developing a new drug was close to €1 billion, in 2015 the estimated cost was €2.58 billion, a 150% increase. Consequently, the different actors of the pharmaceutical R&D industry have an urgent need to improve their effectiveness to develop new drugs. In this sense, the early stage of the R&D process shows the greatest potential to increase the success rate of new drug candidates, as it is in this phase where candidate molecules are selected for further testing in humans. Computer Aided Drug Design (CADD) is the most cost-effective tool for drug discovery. These technologies allow researchers to screen large databases of molecules and simulate its interaction in vivo. However, they are extremely intensive in computational resources. As a result, existing solutions have to implement simplifications in the simulations in order to incur acceptable computational times. These simplifications come at the cost of accuracy, finding a low rate of bioactive molecules that, in addition, usually fail during clinical trials. Pharmacelera aims at creating a new computing standard for Computer-Aided Drug Design (CADD) through an open machine learning platform that enables accurate prediction of molecular properties of candidate molecules based on quantum-mechanics (QM) calculations during drug development. By providing early stage ADME-Tox analysis, thus dramatically reducing the failure rate of the drug design process. In addition, by combining HPC, new disruptive molecular modelling techniques (computational chemistry) and machine learning, PharmAgile will be 20x faster and 10x more precise than any existing solution. Consequently, PharmAgile will be able to reduce the overall costs of drug development by up to 19% and the drug development time up to 8%. This translates into $335 million and 2 years saved per new approved drug.
Fields of science (EuroSciVoc)
CORDIS classifies projects with EuroSciVoc, a multilingual taxonomy of fields of science, through a semi-automatic process based on NLP techniques.
CORDIS classifies projects with EuroSciVoc, a multilingual taxonomy of fields of science, through a semi-automatic process based on NLP techniques.
- medical and health sciencesbasic medicinepharmacology and pharmacydrug discovery
- medical and health sciencesbasic medicinepharmacology and pharmacypharmaceutical drugs
- medical and health sciencesbasic medicinemedicinal chemistry
- medical and health sciencesclinical medicine
- natural sciencescomputer and information sciencesartificial intelligencemachine learning
You need to log in or register to use this function
Programme(s)
Funding Scheme
SME-1 - SME instrument phase 1Coordinator
08173 Saint Cugat Del Valles Barcelona
Spain
The organization defined itself as SME (small and medium-sized enterprise) at the time the Grant Agreement was signed.