Skip to main content
Fig. 3 | Microbiome

Fig. 3

From: MetaLonDA: a flexible R package for identifying time intervals of differentially abundant features in metagenomic longitudinal studies

Fig. 3

Pattern and performance evaluation of data simulated from various statistical distributions. (a) The pattern of the simulated longitudinal features. Each differentially abundant feature has time intervals between group A and B at [1,5] \(\bigcup \) [8,13] \(\bigcup \) [15,20] and non-differential time intervals [5,8] \(\bigcup \) [13,15]. The simulated data mimics inconsistencies in sample collection (different number of subjects per group, different number of samples per subject, and samples not equally spaced.) (b) The fitted smoothing spline of each group and the highlighted significant time intervals between the two groups. (c–e) The performance of different tools using data simulated from NB, Poisson, and zero-inflated Poisson, respectively. Each bar represents the mean among 1000 features, and the error bar represents the standard deviation. MetaLonDA always has a higher specificity than LOWESS and MetaSplines. This shows MetaLonDA’s robustness among different distributions

Back to article page