Fig. 3: Forecasting of the signals.
From: Forecasting the dynamics of a complex microbial community using integrated meta-omics
The 17 signals are predicted for the years 2011–2016 and compared with the data from June for those years. The green and blue dots represent the training and test data, respectively; the solid line depicts the median of the prediction, while the shaded area represents the 95% confidence interval. The green and blue boxplots on the right of every box depict the distribution of the model residuals from the training and the test sets, respectively. Corresponding scales are provided on the right y axes. The residue displacement from the null distribution was assessed using a Wilcoxon two-sided test (n = 21). The * on top of the boxplot indicates a statistical difference (Benjamini-Yekutieli corrected P < 0.01) between the mean of the residual distribution and 0, indicating incorrect/incomplete modelling (exact P values in Supplementary Table 7). In the boxplots, the central line indicates the second quartile, the lower and upper hinges correspond to the first and third quartiles and the whiskers extend from the hinge to the smallest/largest value no further than ±1.5 × the distance between the first and third quartiles. The samples beyond the range are plotted as individual outlier dots.
