Interpretable deep learning uncovers cellular properties in label-free live cell images that are predictive of highly metastatic melanoma

Deep learning hasemerged as the technique of choice for identifying hidden patterns in cell imaging data but is often criticized as black box.'' Here, we employ a generative neural network in combination with supervised machine learning to classify patient-derived melanoma xenografts as efficient'' or inefficient'' metastatic, validate predictions regarding melanoma cell lines with unknown metastatic efficiency in mouse xenografts, and use the network to generate in silico cell images that amplify the critical predictive cell properties. These exaggerated images unveiled pseudopodial extensions and increased light scattering as hallmark properties of metastatic cells. Wevalidated this interpretation using live cells spontaneously transitioning between states indicative of low and high metastatic efficiency. This study illustrates how the application of artificial intelligence can support the identification of cellular properties that are predictive of complex phenotypes and integrated cell functions but are too subtle to be identified in the raw imagery by a human expert. A record of this paper's transparent peer review process is included in the supplemental information.

Automated summary

The study utilized a generative neural network and supervised machine learning to classify the metastatic efficiency of melanoma cells, revealing characteristic features of metastatic cells and validating these findings with live cell observations. It demonstrates how artificial intelligence can aid in identifying cell properties predictive of complex phenotypes and integrated cell functions that may be subtle and difficult for human experts to identify in raw imagery.