Fig. 1From: Machine learning analysis of microbial flow cytometry data from nanoparticles, antibiotics and carbon sources perturbed anaerobic microbiomesUnsupervised autoencoder analysis can be used to identify significantly perturbed microbiomes. The mean squared error (MSE) between actual value and reconstruction has been displayed on the y-axis for each sample tested. The red horizontal line at 0.05 MSE represents a threshold error to decide an outlierBack to article page