Fig. 2
From: Disentangling environmental effects in microbial association networks
Quantification of environmentally driven associations in the BBMO network A The first column shows the number (in thousands, K) or fraction of microbial associations divided by domain: bacteria–bacteria associations (B), bacteria–eukaryote associations (BE), and eukaryote–eukaryote associations (E). The second column shows the number (or fraction) of associations divided by size-fractions: association within the nano size fraction (n), within the pico size fraction (p), and between these two size fractions (np). The third column shows all microbial edges connected to an environmental parameter: temperature (Tem), day length (day), chlorophyll (Chl), inorganic nutrients NO3− (NO3), SiO2 (Si), and NO2− (NO2). The last column shows the number of edges divided in how many triplets they have been found ranging from no triplets (0) to six triplets. The first two rows display the number of microbial associations of the BBMO network before applying EnDED. Positive associations are indicated with black, negative associations with red. The last two rows indicate in blue the fraction of environmentally driven edges among the positive (3rd row) and negative (4th row) microbial associations. B The left network shows in black the positive and in red the negative associations. The right network shows the number of triplets a microbial edge is in ranging from one (green) to six (orange), and no triplet (black). The middle network shows in blue the environmentally driven associations that were detected by the intersection combination of the four methods sign pattern, overlap, interaction information, and data processing inequality