Computational Structural Bioinformatics Workshop 2022 (CSBW 2022)

December, 7th 2022, Las Vegas, NV

Understanding the fundamental role of protein and nucleotide structures in life processes involves many computational challenges and opportunities. New experimental and computational techniques that generate and use large amounts of structural data have enabled novel analysis approaches, leading to an increased understanding of different biophysical phenomena. The Computational Structural Bioinformatics Workshop aims to provide a forum for discussion of state-of-the-art computational approaches related to protein structure discovery and analysis. Program at: https://facultyweb.cs.wwu.edu/~jagodzf/csbw/programs/CSBW2022.pdf

Panel Discussion on AI and ML in Structural Biology

The panel discussion included panelists Prof. Jing He, from Old Dominion University and Prof. Filip Jagodzinski from Western Washington University. There was a lively discussion on the current role of artificial intelligence and machine learning on current developments in computational structural biology. Of note, audience members were interested in how graduate students are trained in the field. Prof. Jagodzinski noted that "We have both undergraduate and Master's level students who are successful in machine learning […] Although I do have to say, the reality is to some extent undergrads and Master's level have to look at those things as black boxes to some extent." There was also discussion on how this is a very fast moving field, and Prof. He mentioned that "I really hope Ph.D students will not be satisfied with the level just to be able to use a tool […]. Try to go in deep, even if small you understand the science part of it […]. I think [this is] the gap that needs to be filled. We are scientists we're not just engineering to apply some of the framework." Prof. Jagodzinski added: "One thing that I've heard from students that come back, who have been successful, they attribute their success to their flexibility and their ability to pick up a new tool quickly and apply to whatever task. It's not just one thing that they're learning about, it's not just the syntax of pytorch or whatnot. They're learning concepts." The role of academia was also discussed in the panel, as many labs now compete, in direct and indirect ways with a well-funded, high power AI industry. Prof. Jagodzinski said "Faculty have an advantage that we can explore different avenues without too many repercussions of going down a bad path. I think if we embrace that, if we instill that in students, I think that opens up doors and possibilities. I think that's a really good mindset for students to have as well." Audience members also stressed the importance of collaborations in these scenarios. Prof. Jagodzinski agreed and added: "The fact that scooping is a phenomenon means that it's a very young discipline or or a topic. So from my experience, facilitating these sort of interdisciplinary cross-disciplinary projects with biologists and chemists, I think there's a higher chance that two people or even more, with different perspectives, will be able to identify a problem that nobody else has identified and then you're the first to attack it." In terms of collaboration, it was also discussed how to communicate across different fields, and making collaborators understand that machine learning cannot give the answer to every problem. Prof. He mentioned she didn't like machine learning as a black box problem and that people can work together to explain what happens behind the scenes, so that others can understand the abilities of machine learning. Prof. Jagodzinski pointed out that it is always better to start from a specific problem, instead of starting from tool creation. The panelists were asked about their view of the field in a few years from now. Dr. He believes there will be more AI in structural bioinformatics but she wants to see more in-depth discoveries from these models. Prof. Jagodzinski stressed big data as the main challenge in upcoming years. Finally, a student from the audience asked how the panelists motivate their students. Prof. He mentioned "ultimately, it's the desire to discover". Prof. Jagodzinski mentioned that studying something that is new can be scary but is also exciting for students.

Videos:

Funding Information:

Funded by the National Science Foundation: Award number IIS-2231497.