Skip to main content
Fig. 4 | Microbiome

Fig. 4

From: An expectation-maximization algorithm enables accurate ecological modeling using longitudinal microbiome sequencing data

Fig. 4

BEEM analysis of year-long gut microbial time series datasets. a, b BEEM-estimated biomass values for two individuals (DA and M3) with daily sampled, year-long gut microbial time series datasets from David et al. [42] and Caporaso et al. [41]. Interestingly, while M3’s biomass fluctuates rapidly, DA’s biomass seems to vary in a more defined fashion with a periodicity of around 3 months. c, d Graphs representing non-zero interaction terms in gLVM models learnt individually for DA and M3 using BEEM. Dashed and solid edges represent positive and negative interactions, respectively. Edge widths are proportional to the interaction strength, and node sizes are proportional to the log-transformed mean relative abundance of the corresponding species. Nodes are labeled with the most specific taxonomic annotations and colored according to order level information

Back to article page