FYIM 2025 Theme and Discussion Questions

Do we still need “bottom-up” mechanistic models and modelers in cell biology?

If the answer is yes, then how best to leverage new AI tools? At what step of the modeling process will AI be most valuable? Or, is there going to be a “doomsday” (depending on your perspective), when AI takes over the modeling field? If so, what are the roles of modeling, modelers, and experimentalists in the meantime? Can AI build mechanistic models? Or will it be sufficient to have a black box, and mechanistic models won't even be relevant or necessary?

Where best to focus large-scale investment? (funding, consortia, etc.)

What kinds of data types and resources will we need to train AI? Which data initiatives do we (the cell-bio-modeling community) recommend for investment, given limitations in funding? Is it possible to build a general AI platform to make modeling easier for members of the field?

What can we learn from computational advances in neighboring fields? 

For example, there has been recent progress in Physics-informed Machine Learning, which enforces physical laws to improve learning algorithms1. The concept has already been extended to Biologically-informed neural networks2. What does the landscape look like for emerging techniques that marry the power of AI/ML with traditional mechanistic modeling in biology? 

Another field to watch is that of protein structure prediction, where a deep, mature computational research field has now merged with excellent data resources and AI, celebrated by a Chemistry Nobel Prize. Bowman argues that3, nevertheless, the protein structure prediction problem is not “solved”. 

1.         Karniadakis, G. E. et al. Physics-informed machine learning. Nat. Rev. Phys. 3, 422–440 (2021).

2.         Lagergren, J. H., Nardini, J. T., Baker, R. E., Simpson, M. J. & Flores, K. B. Biologically-informed neural networks guide mechanistic modeling from sparse experimental data. PLoS Comput. Biol. 16, e1008462 (2020).

3.         Bowman, G. R. AlphaFold and Protein Folding: Not Dead Yet! The Frontier Is Conformational Ensembles. Annu. Rev. Biomed. Data Sci. 7, 51–57 (2024).