Machine learning and AI can offer a shortcut to more effective and safer compounds and pharmaceuticals for human health
Artificial intelligence-based methods are effective tools for the analysis of big data and to identify patterns in them. They can be used to aid chemical safety assessment and to discover drugs that exert specific molecular alterations and therapeutic effects. Senior Research Fellow Angela Serra believes that computational approaches can be used to develop more effective and safer therapeutic strategies for human health.
Italian computer scientist Angela Serra is passionate about how machine learning and artificial intelligence algorithms can be used to analyze big data and solve complex biological problems. She specialized in integrated learning approaches for biomedical data during her PhD and moved to Tampere as a postdoctoral researcher in 2018. Currently she is a Senior Research Fellow at the Finnish Hub for Development and Validation of Integrated Approaches, where she aims to develop computational strategies for drug discovery and chemical safety assessment.
Artificial intelligence accelerates the development of compounds and their safety evaluation
According to Serra, artificial intelligence methods have a lot of potential in the development process of drugs and chemical compounds.
– The purpose of my research is to develop computational models for the analysis of omics and chemical data. Classical cheminformatic methods are used to link the chemical structure of a compound with its effects. While these methods provide important insight into relevant chemical properties, they might fail to explain the underlying biological mechanism. Toxicogenomics makes use of omics technologies to characterize and study the effects of chemicals or pharmaceuticals on human health. The development of novel computational methods that integrates the strength of both cheminformatic and toxicogenomic methods can aid and speed-up the design of new chemical compounds with desired effects, Serra explains.
For example, in the drug development process, the drug discovery phase aims at identifying new candidate medication. Currently, the process is expensive and time consuming because it involves screening thousands of compounds to identify one or a few molecules with the desired efficacy. Machine learning and artificial intelligence algorithms can significantly speed up the process, while keeping it reliable and safe
A more efficient development process makes chemistry more sustainable
In recent years, Serra has developed multiple tools for analyzing and interpreting large data sets for the characterization of the mechanisms of action of compounds and their safety assessment in an unbiased and reliable manner. Chemical safety assessment aims to uncover the risks of chemical exposures and to define limits for the safe use of the compounds.
– These tools can aid the discovery of effective compounds and their safety assessment process. With respect to the drug development, if issues like safety are neglected at the initial stages of the process, there can be higher probability of failure of clinical trials, leading to wasted effort, time, and money, Serra says.
– In the future I want to bring my research to the next level by developing new AI-based strategies that will help optimize the discovery of chemicals at early development stages. I aim to reduce the number of experimental tests, hence reducing the environmental impact of the process while promoting greener chemistry, she concludes.
Keywords: AI, artificial intelligence, machine learning, computational science, pharmaceutical drugs, drug development, chemistry