Heterogeneity in cellular phenotype is a key biological determinant of biological functions and is challenging to identify. To tackle this problem, we developed a deep learning method called AINU (artificial intelligence [AI] of the nucleus) that recognizes specific nuclear signatures. AINU can identify cellular heterogeneity and differentiate between various cell states using a small number of super-resolution microscopy data. AI interpretability methods showed that RNA Polymerase II localizations in the nucleoli are a discriminator of cell state.
[Read More...]