By Nancy Pomarici
The conference featured various sessions, which included symposia, platform talks and lectures. The symposium Predicting Protein Fold was particularly interesting, since this has become a major field of research and outstanding progress was made in the last years. I especially enjoyed the talk of Dr. Kathryn Tunyasuvunakool, who is today part of the team behind DeepMind’s AlphaFold, a protein-structure-prediction tool. AlphaFold is a machine-learning model that can predict a protein’s structure from its amino-acid sequence. In the last two years, this tool represented a breakthrough in the field, performing much better than any other existing method. In her talk, Dr. Tunyasuvunakool demonstrated how AI models can be used to make sense of complex data and to generate hypotheses. After the talk, an interesting discussion followed. A major point raised by some scientists was that, even if AlphaFold showed an extraordinary performance in the last Critical Assessment of Structure Prediction (CASP) competitions, it sometimes fails when applied in everyday scientific worklife. The reason behind is that, even if the tool is easy to use, it requires a lot of expertise and knowledge about the topic, and a deep scientific criticism is necessary to prove the validity of the result; in general, the protein structure prediction field made huge advances, but further improvements are still required.
AI was a central topic also in many other talks and lectures. The Associate Professor Pratyush Tiwary, from University of Maryland, is an expert in modelling rare events, using molecular dynamics simulation and Artificial Intelligence. During the BPS, I had the honor to attend his lecture, in which he described how AI can help molecular dynamics simulations to study processes that are too slow even for the best supercomputer. His talk raised my interest in these new techniques, which I will study more in details and then try to apply to my open questions.
Attending the conference gave me the opportunity to connect with several researchers and scientists who shared similar research interests. I was able to exchange ideas and discuss potential collaborations with a few of them during the coffee breaks and poster sessions. In the last day of the conference, I also presented my poster about stability of monoclonal antibodies, and this was a memorable experience. Many people showed interest in my work, from students to professors. I especially enjoyed talking to the employees of some pharma companies: they gave me insights about my research from a completely new point of view, opening questions that can become new research areas. Another topic of our conversation was also the comparison of academic research to the one performed in a pharma company, finding out some similarities, but also many divergent aspects. This discussion was helpful for me to get to know experiences of people coming from different environments, opening my mind about different realities, and inspiring my future research life.
The conference provided me with several suggestions and ideas that I plan to implement in my future work. One of the key takeaways was the importance of multi-disciplinary collaborations in research. Especially for theoretical research, as in my case, the collaboration with experimentalists is often key to reach innovative solutions for complex problems.
Overall, attending the Biophysical Society Meeting was an extremely enriching experience. It provided me with the opportunity to learn about the latest developments in computational techniques and protein research, connect with fellow researchers and experts, and gain valuable insights for my future work. I am confident that the knowledge and suggestions obtained from the conference will help me in my career and contribute to the advancement of the field.