Neural network shows affinity for predicting properties of organic materials

Researchers at the Tokyo Institute of Technology (Tokyo Tech) in Japan have used artificial intelligence (AI) to predict the degree of water repulsion and protein adsorption by ultra-thin organic materials. By making accurate predictions of water repulsion and protein adsorption for even hypothetical materials, the researchers' approach, reported in a paper in ACS Biomaterials Science & Engineering, opens up new possibilities for the screening and design of organic materials with desired functions.

The use of informatics in the field of inorganic material design has led to the development of new types of catalysts, batteries and semiconductors. In contrast, the informatics-based design of biomaterials (i.e. organic rather than inorganic solid-state materials) is only just beginning to be explored.

A team of researchers at Tokyo Tech led by associate professor Tomohiro Hayashi has now successfully made inroads into this emerging field. They used machine learning with an artificial neural network (ANN) to predict two key properties – the degree of water repulsion and the affinity with protein molecules – of ultra-thin organic materials known as self-assembled monolayers (SAMs). Due to their ease of preparation and versatility, SAMs have been widely used to create model organic surfaces for exploring the interaction between proteins and materials.

Using a literature-based database of 145 SAMs, the researchers trained the ANN to be able to predict water repulsion (measured in terms of the water contact angle) and protein adsorption accurately. They then went on to demonstrate that the ANN could predict water repulsion and protein adsorption even for hypothetical SAMs.

SAMs are attractive for the development of many applications in organic electronics and the biomedical field. The two properties investigated in this study are of enormous interest to biomedical engineers.

"For example, implant materials that exhibit low water contact angle enable fast integration with the surrounding hard tissues," Hayashi says. "In the case of artificial blood vessels, the resistance to the adsorption of blood proteins, in particular fibrinogen, is a critical factor to prevent platelet adhesion and blood clotting."

Overall, the study opens the door to advanced material screening and design of SAMs but with potentially greatly reduced costs and time scales. The researchers plan to continue scaling up their database and, within a few years, to expand their approach to include polymers, ceramics and metals.

This story is adapted from material from the Tokyo Institute of Technology, with editorial changes made by Materials Today. The views expressed in this article do not necessarily represent those of Elsevier. Link to original source.