Net Surfing

Successfully completed the project ATTRACTOR

Progetto Attractor

Successfully completed the project ATTRACTOR

At the conclusion of the feasibility study strumenti di intelligenza ArTificiale per oTtimizzare il RilAscio ConTrollatO di faRmaci(ATTRACTOR), the realization of the Proof of Concept ATTRACTOR based on AI was successfully completed.

The feasibility study carried out with the partner Aethia and in collaboration with the University of Turin (dept. Chemistry) in the field of precision medicine focused on analysing and identifying different materials (carriers) suitable for the release of drugs into the body. Simulating the behaviour of the different carriers using quantum-mechanical methods is very expensive, while machine learning (ML) is a valid alternative.

The PoC realized with Machine Learning algorithms could be a great help for a pharmaceutical company or research centre, to study and prepare suitable carriers without running all the simulations. In fact, with limited times and costs, the search field is narrowed down to a smaller number of candidates on which to concentrate more in-depth research at a later time.

The feasibility study characterized a platform suitable for the needs and dimensions of an average professional reality and for evaluating the necessary application software in terms of hardware, software and network resources.

In carrying out the feasibility study, at least two ML models were developed, albeit in a prototype way, capable of making predictions on the structure-property relationships of metal-organic microporous materials (metal-organic frameworks, MOFs).

In fact, MOFs have the characteristic of being chemically very versatile because they are made up of structural units that can be modulated and assembled with different topologies, like in a “molecular lego”. Furthermore, the three-dimensional porous lattices are characterized by cages and channels of different shapes and sizes which make them selective for the capture and release of molecules.

The results obtained with MOF molecules have been verified for which the distribution of electric charges has been accurately calculated and the results obtained with the predictive model: the accuracy is impressive and a Web App is now available which allows the MOF structures to be subjected to models of ML and returns the prediction of the stability of the structure itself.

The web App is accessible at the following URL:


Progetto in sviluppo con fondi di investimento europei 2014/2020

Obiettivo tematico I – Ricerca, sviluppo tecnologico e innovazione Azione I.1.b.1.2 “Sostegno alle attività collaborative di R&S per lo sviluppo di nuove tecnologie sostenibili, di nuovi prodotti e servizi – Bando PASS