Lugo-Martinez, Jose Ruiz-Perez, Daniel Narasimhan, Giri Bar-Joseph, Ziv Additional file 7 of Dynamic interaction network inference from longitudinal microbiome data Figure S5. Learned dynamic Bayesian network for gut and vaginal microbiomes derived from unaligned samples. Figure shows two consecutive time slices ti (orange) and ti+1 (blue), where nodes are either microbial taxa (circles) or clinical/demographic factors (diamonds). Nodes size is proportional to in-degree whereas taxa nodes transparency indicates mean abundance. Additionally, dotted lines denote intra edges (i.e., directed links between nodes in same time slice) whereas solid lines denote inter edges (i.e., directed links between nodes in different time slices). Edge color indicates positive (green) or negative (red) temporal influence, and edge transparency indicates strength of bootstrap support. Edge thickness indicates statistical influence of regression coefficient as described in network visualization. a Learned DBN for the unaligned infant gut microbiome data at a sampling rate of 3 days and maxParents = 3. b Learned DBN for the unaligned vaginal microbiome data at a sampling rate of 3 days and maxParents = 3. (PDF 56 kb) Dynamic interaction network inference;Longitudinal microbiome analysis;Microbial composition prediction;Dynamic Bayesian networks;Temporal alignment 2019-04-02
    https://springernature.figshare.com/articles/journal_contribution/Additional_file_7_of_Dynamic_interaction_network_inference_from_longitudinal_microbiome_data/7943390
10.6084/m9.figshare.7943390.v1