StressNET: an adaptable deep-learning model for mechanical stress inference in tissues
Mechanical interactions between cells are fundamental to tissue morphogenesis during development and regeneration. Computational methods that infer intercellular stresses from microscopy images of cell shapes offer a non-invasive alternative to experimental perturbation techniques, yet all existing approaches rely on explicit physical models. Here we present StressNET, a Graph Neural Network…
Mechanical interactions between cells play a crucial role in the development and regeneration of tissues. While computational methods exist to infer intercellular stresses from microscopy images of cell shapes, all current approaches rely on explicit physical models. Researchers have now introduced StressNET, a Graph Neural Network (GNN) that infers these stresses directly from tissue geometry, without the need for any underlying physical model.
To train and benchmark StressNET, the team generated synthetic datasets. The findings show that StressNET's predictions correlate highly with experimental stress proxies in zebrafish neuromasts and Xenopus embryos. Upon analyzing the network's latent space, the researchers discovered that StressNET learns global organizational principles of mechanical stress distribution, going beyond just local cell-cell interactions.
The researchers made StressNET open-source and provided pre-trained models that can be fine-tuned on in vivo data, making it a highly adaptable tool for studying tissue mechanics across various biological systems.
Written by urgent.news from bioRxiv's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
