No Foundations without Foundations -- Why semi-mechanistic models are essential for regulatory biology

TL;DR

This paper advocates for semi-mechanistic models that integrate mechanistic insights with experimental design to advance predictive understanding in regulatory biology, emphasizing the importance of first-principles approaches over purely data-driven methods.

Contribution

The authors introduce a semi-mechanistic framework unifying perturbation experiments across various systems, linking it to existing models and demonstrating improved predictive performance.

Findings

Modified loss function enhances prediction accuracy

Error analysis informs batching strategies

Framework clarifies assumptions in machine learning methods

Abstract

Despite substantial efforts, deep learning has not yet delivered a transformative impact on elucidating regulatory biology, particularly in the realm of predicting gene expression profiles. Here, we argue that genuine "foundation models" of regulatory biology will remain out of reach unless guided by frameworks that integrate mechanistic insight with principled experimental design. We present one such ground-up, semi-mechanistic framework that unifies perturbation-based experimental designs across both in vitro and in vivo CRISPR screens, accounting for differentiating and non-differentiating cellular systems. By revealing previously unrecognised assumptions in published machine learning methods, our approach clarifies links with popular techniques such as variational autoencoders and structural causal models. In practice, this framework suggests a modified loss function that we demonstrate can improve predictive performance, and further suggests an error analysis that informs batching strategies. Ultimately, since cellular regulation emerges from innumerable interactions amongst largely uncharted molecular components, we contend that systems-level understanding cannot be achieved through structural biology alone. Instead, we argue that real progress will require a first-principles perspective on how experiments capture biological phenomena, how data are generated, and how these processes can be reflected in more faithful modelling architectures.