Fig. 2: Summary of the prediction accuracy of each cluster from the 24 WWTP test datasets. | Nature Communications

Fig. 2: Summary of the prediction accuracy of each cluster from the 24 WWTP test datasets.

From: Predicting microbial community structure and temporal dynamics by using graph neural network models

Fig. 2: Summary of the prediction accuracy of each cluster from the 24 WWTP test datasets.

The values are calculated using the Bray-Curtis dissimilarity measure (0.0 means identical and 1.0 completely different) of each cluster (group of ASVs). The different clustering types are indicated by colors according to the legend. The x-axis has been square-root transformed to better reveal differences at the lower values. Data are presented as median, 25th–75th percentiles, and whiskers extending to 1.5×IQR; outliers shown as points. The datasets are ordered by the number of samples in each dataset, which is indicated in parentheses. Source data is provided as a Source Data Fig. 2.

Back to article page