Deep Learning Neural Networks Highly Predict Very Early Onset of Pluripotent Stem Cell Differentiation

WAISMAN, ARIEL; LA GRECA, ALEJANDRO; MÖBBS, ALAN M.; SCARAFÍA, MARÍA AGUSTINA; SANTÍN VELAZQUE, NATALIA L.; NEIMAN, GABRIEL; MORO, LUCÍA N.; LUZZANI, CARLOS; SEVLEVER, GUSTAVO E.; GUBERMAN, ALEJANDRA S.; MIRIUKA, SANTIAGO G.

Stem Cell Reports

Cell Press

Año: 2019 vol. 12 p. 845 - 859

2213-6711

Deep learning is a significant step forward for developing autonomous tasks. One of its branches, computer vision, allows image recognition with high accuracy thanks to the use of convolutional neural networks (CNNs). Our goal was to train a CNN with transmitted light microscopy images to distinguish pluripotent stem cells from early differentiating cells. We induced differentiation of mouse embryonic stem cells to epiblast-like cells and took images at several time points from the initial stimulus. We found that the networks can be trained to recognize undifferentiated cells from differentiating cells with an accuracy higher than 99%. Successful prediction started just 20 min after the onset of differentiation. Furthermore, CNNs displayed great performance in several similar pluripotent stem cell (PSC) settings, including mesoderm differentiation in human induced PSCs. Accurate cellular morphology recognition in a simple microscopic set up may have a significant impact on how cell assays are performed in the near future. In this article, Miriuka and colleagues show that deep learning convolutional neural networks can be trained to accurately classify light microscopy images of pluripotent stem cells from those of early differentiating cells, only minutes after the differentiation stimulus. These algorithms thus provide novel tools to quantitatively characterize subtle changes in cell morphology.