Waddington Landscape as a “colorful metaphor” is often used to describe the transition of cell fates. Cells are located at various locations in the Waddington landscape, corresponding to multiple differentiation potentials. Conventionally, characterizing cellular differentiation potency relies on transcriptomic signatures of known markers. However, it is not enough to reflect the global molecular characteristics of cells based on just a few markers. Traditional marker-based methods are also not suitable for identifying novel cell populations.

To address this conundrum, the research team from Inner Mongolia University has proposed a new energy indicator based on a feedback neural network - Hopfield neural network (HNN), which can quantitatively characterize the cellular differentiation potency and Waddington landscape without prior knowledge.

Schematic diagram of Waddington's energy landscape in embryogenesis (left) and cell reprogramming (right).

The researchers quantitatively evaluate the differentiation potency of different stem cells and indicated that cellular differentiation potency can be approximated by Hopfield energy values. Given the potential relationship between Hopfield energy and cell differentiation potential, the researchers profiled Waddington's energy landscape during the embryogenesis and reprogramming process containing explicit time information. The energy landscape at single-cell resolution further confirmed that cell fate decision is progressively specified in a continuous process. Finally, the researchers proposed an importance score (IS) to identify the key factors driving cell fate transitions, providing an effective strategy for purposefully controlling the cell fate transition.

This study is an effort to move the qualitative Waddington landscape toward a quantitative model, which provides a general measure of differentiation potency independent of external knowledge.

Tag: Health Science

sources: https://spj.science.org/doi/10.34133/research.0118