Fig. 4

Co-occurrence analysis of features from green inter-rows. a Network analysis of core ASVs from aCO₂ rings and b eCO₂ rings. Features with an absolute ALDEx2 effect size > 0.5 were utilized for SpiecEasi analysis applying the Meinshausen & Bühlmann (mb) method with a number of subsamples of 50, n-lambda of 100 and lambda minimum ratio of 0.1; blue and red edges indicate positive and negative co-occurrence respectively; size of the nodes is proportional to the number of ASV reads. Partial correlation analysis of genera with an absolute ALDEx2 effect size > 0.1 from c aCO₂ and d eCO₂ green inter-rows using SpiecEasi and SPRING. SpiecEasi run applying the Meinshausen & Bühlmann (mb) method and SPRING with a modified centered log ratio (mclr). Both analyses utilized a number of subsamples of 99, a lambda minimum ratio of 0.1 and the Stability Approach to Regularization Selection (StARS) using co-occurences with a threshold of < -0.5 and > 0.5