ATTRACTOR
Feasibility study in the field of precision medicine, focused on the analysis and identification of different materials (carriers) suitable for drug delivery in the body.
Description
Feasibility study in the field of precision medicine, focused on the analysis and identification of different materials (carriers) suitable for drug delivery in the body. Simulating the behaviour of different carriers using quantum-mechanical methods is very time-consuming and costly, but machine learning (ML) is a viable alternative. The feasibility study ended up in the realization of a platform suited to the needs and size of an average professional organisation, and to the evaluation of the application software required in terms of hardware, software and network resources. The feasibility study achieved its objectives more accurately than expected, and a Web-App that allows a pharmaceutical company or research centres to study and prepare suitable carriers without running all the simulations is now available.
The Web-App can be accessed at the following URL: http://attractor.netsurf.it/
Result to be enhanced
Two Machine Learning models have been developed with different technologies. The models have brought similar results, albeit with different performances and similar precision. In particular, model with open source XGBurst technology, new model created with TensorFlow and neural network, more dynamic and with excellent results. For both models, the error between calculated data compared to the predicted ones is absolutely acceptable, with great satisfaction for both models created. For this reason, it has been made possible for the user who requests a prediction to choose which model to use.
The success of the results obtained gives reason to create new opportunities for the creation of a professional service based on the models created, which can be proposed on the market for research centers that intend to select MOF molecules suitable for completing computational analysis, with the aim of creating drug ‘carriers’ for “precision medicine”.
Why is it important?
It could be a valuable help for pharmaceutical companies and research centres, which could use it to study and prepare suitable carriers without running all the simulations. In fact, with limited time and costs, one could narrow down the research field to a smaller number of candidates on which to focus a more in-depth research at a later stage. The business hypothesis linked to the success of this feasibility study is based at first on the possibility of offering a service to support the targeting of analyses for research centres, and subsequently becoming a service portal for pharmaceutical companies.
Project and Acronym: ATTRACTOR
TRL: Iniziale 2 – Finale 4
Reference call: PASS
Innovation Cluster to contact: Polo ICT
Technologies used: Machine learning, deep learning
Lead company:
Net Surfing S.r.l
enrico.baratono@netsurf.it | www.netsurf.it
Collaborating company:
Aethia S.r.l