Finding signatures of low-dimensional geometric landscapes in high-dimensional cell fate transitions

The preprint operates within the paradigm of cell fate transitions being governed by underlying, low-dimensional landscapes, often described through the metaphor of Waddington's epigenetic landscape. It also draws on the related paradigm of dynamical systems theory, specifically focusing on the concept of attractor states and their bifurcations as a means of characterizing cell fate decisions.

This preprint presents a model that integrates the established paradigms of Waddington's landscape and Hopfield networks, aiming to bridge the gap between theoretical descriptions of cell fate transitions and high-dimensional gene expression data. It leverages established concepts within these paradigms, like attractor states, bifurcations, and order parameters, and introduces modifications, like signal-dependent potentials and generalized order parameters, to enhance the model's applicability to real biological systems. It applies the model to existing scRNA-seq datasets, demonstrating its ability to identify signatures of different decision-making classes within the paradigm of dynamical systems theory. While the model introduces some novel elements, its core relies on established theoretical frameworks and primarily extends their application, therefore, the preprint resides between the Normal Science and Model Drift stages. I have chosen Normal Science due to its demonstration of the utility of existing theoretical constructs to analyze new data, hence its lack of a radical conceptual departure. It uses these paradigms to extract biologically relevant features and connect high-dimensional molecular data to theoretically informative descriptions. The modifications to the Hopfield model to include signal-dependent dynamics and the integration of generalized order parameters represent a refinement and extension of existing tools within the paradigm, pushing towards Model Drift but not fully crossing the threshold to a significant departure from the accepted model.