iModulonDB
Genes
The iModulon weight of all genes is shown. Genes outside of the iModulon thresholds (vertical lines) are considered to belong to the {iModulonName} iModulon.
The genomic position of each gene is shown. Genes above/below the iModulon thresholds (horizontal lines) are considered to belong to the {iModulonName} iModulon.
The activity of the {iModulonName} iModulon is shown across {numProjects} projects in the {datasetName} dataset. To learn more about the experimental conditions of each project, return to the dataset page.
Regulation
Regulons are groups of genes known to be co-regulated by a transcription factor. Regulon genes are often identified using the location of known transcription factor binding sites. By comparing various features of the genes in iModulons and Regulons across the dataset, we can identify differences between top-down, data-driven (iModulon) and bottom-up, bio-molecular (Regulon) approaches to find sets of co-expressed genes.
When an iModulon can be mapped to a known regulon, we can compare:
- The gene overlap between the iModulon and its Regulon (venn diagram).
- The correlation matrix of the gene expression of iModulon and Regulon genes (clustered heatmap).
- The linkage tree of the correlation matrix (dendrogram).
- The correlation between the iModulon activity and the gene expression of its mapped regulator(s) (scatter plot). To compare the strength of this iModulon-Regulator correlation to other iModulon-Regulator pairs, check out the Analysis page.
The venn diagram shows the overlap between the genes identified in the iModulon and the genes in its mapped regulon. Mouse over the graph to reveal the genes in each section.
The clustered heatmap shows the correlations between the expression of genes in the iModulon (yellow boxes), genes in the Regulon (black boxes) and genes in both the iModulon and the Regulon (green boxes).
The heirachical clustering of the correlation matrix (metric = 'euclidean' , method = 'avg') results in the dendrogram to the left. Excluding larger nodes, the size of each node represents the total number of genes in the child nodes. Each node is displayed as a pie chart that reflects the the fraction of genes that belong to the iModulon (yellow), Regulon (black), or both (green). Mouse over the graph to see the gene and gene products present in each node.
iM Activity vs Gene Expression
Using ICA, we find genes that belong to the iModulon. Plotting the expression of each gene against the iModulon activity can help us confirm if a gene belongs in the iModulon. The linear regressions, Pearson r and R2 values are shown. Strong (R2 > 0.64), moderate (0.64 > R2 > 0.16), and weak (0.16 > R2) correlations are depicted by solid, dashed, and dotted lines, respectively. Only experimental conditions that do not reflect pertubations to the iModulon gene are used to calculate the linear correlations. Perturbations are denoted with an X.