Network sector interactions are common in the regulation of the transcriptome response change to flg22 treatment.

<p>Note that even three- and four-sector interactions are common. A heatmap is shown for hierarchical clustering of 5189 flg22-responsive and network-dependent genes and the two hormones SA and JA (rows); clustering is based on their signaling allocations (columns). The signaling allocation model for each gene or hormone at all time points was regularized using lasso and the stringency of regularization was selected based on the AICc. The sign and magnitude of the signaling allocation contributions are indicated by color and intensity; red and blue for positive and negative, respectively; allocations removed by regularization are set to zero—visually they are white. Allocations were calculated for each gene or hormone from the transcript or hormone response changes through the time-series 1, 2, 3, 5, 9, and18 hours post treatment (hpt); 6 minor columns representing the 6 time points post treatment follow each other consecutively within each major column. Individual sector contributions are shown in the first four major columns and are indicated as JA, ET, PAD4, and SA. Sector interactions are shown in the rest of the major columns and indicated using ‘:’ with single letter notations of the sectors, for example the interaction between JA and ET is denoted ‘J:E’. Columns a and b on the right are indicator columns that mark, with black rows, (a) the 2979 genes whose transcript response relative to wild type is buffered in all single-sector mutants (<i>dde2</i>, <i>ein2</i>, <i>pad4</i>, and <i>sid2</i>), and (b) the 1564 genes whose transcript response relative to wild type is buffered in all single and double- sector mutants (<i>dde2</i>, <i>ein2</i>, <i>pad4</i>, <i>sid2</i>, <i>dde2 ein2</i>, <i>dde2 pad4</i>, <i>dde2 sid2</i>, <i>ein2 pad4</i>, <i>ein2 sid2</i>, and <i>pad4 sid2</i>). Red and blue arrows on the left of the heatmap indicate the rows corresponding to free JA and free SA, respectively. The 5189 flg22-responsive and network-dependent genes were selected according to the procedure described in the Materials and Methods. Signaling allocation rows were scaled for better visualization using the most extreme value of each row.