Decoding gut microbiota by imaging analysis of fecal samples

In this study, we demonstrate that the deep neural network approach using low-resolution microscope images of fecal gram staining samples enables us to characterize the dynamic changes in murine gut microbiota. We obtained images of fecal samples from high-fat diet treated obese mice and dextran sulfate sodium-treated colitis mice and applied them to convolution neural network.