This tab displays the activity and activity fold change of up to 20 TFs over each cluster in the selected brain region.
• Clusters are ordered according to the dendrogram which represents a molecular taxonomy of all cell populations
• Below the dendrogram, clusters are annotated by brain region, and time point.
• Bubble colour encodes the mean TF activity within the cluster, and bubble size encodes the fold change of TF activity in each cluster compared to all other clusters - effectively describing TF specificity.
• The "Scaling" option allows the option of normalizing mean TF activity of all TF from 0, 1 for a better visualization when different TFs have large differences in absolute AUC value. This does not change the FC values.
• Hover over each bubble to get additional details about each cluster & its expression level
This tab displays a heatmap of user selected TF activity per cluster
• Values in the heatmap represent the mean TF activity per cluster.
• Joint clusters are classified based on the combined data from every developmental time-point in a brain region (forebrain or pons); sample cluster are identified based on data from each individual time-point per brain region.
• Use the "Time-point to Visualise" option to select which (if not all) time-points to visualise in the sample cluster heatmap.
This tab displays the activity of selected transcription factors
• Cells are plotted in 2D according to selected dimensionality reduction algorithm
• The top row is colored by joint cluster and the bottom row is colored by transcription factor activity
• Only the first two transcriptions factors are displayed
This tab displays a network visualisation of the inferred regulatory relationship between TFs and target genes.
• TFs and target genes are represented as nodes with regulatory relationships represented as edges.
• User input TFs are shown in blue. Target genes that are present in the current network can be highlighted in orange based on a user-input gene list, either through the "Genes of Interest"input or through a file input.
• Click on the "Label Target Gene Nodes" option to see the label of every gene target and enable a hover option. Currently, hover only displays gene name but more information to come soon!
• TFs that self regulate are not displayed (i.e no self loops).
• File input format: single column csv file with the first row titled 'Gene' and the remaining rows containing a list of genes of interest.
Network Visualisation Download (PDF)
This tab displays information corresponding to the selected TFs and their inferred target genes.
• Strength of Association represents the weight of the putative regulatory links between transcription factor and a gene target, as predicted with Genie3, with a higher value indicating a more likely regulatory link.
• The number of motifs for each gene is identified via the RcisTarget package. The best motif and its sequence logo is displayed.
This plot quantifies the proportion of cells (from 0 to 1) at each timepoint where a given TF is active, broken down by cell type, to allow for visualizing activity across time.
• For any given cell, any given TF is considered active if its activity in that cell is higher than a TF activity threshold determined in the SCENIC pipeline.
• The time series for the first TF selected in the sidebar will be an interactive plot, with the remaining plots being static.
Interactive Plot Options: double clicking a cell type in the legend displays that cell type ONLY; single click removes that cell type from the plot. Mouse over ribbons in the plot to see the cell types. We only support four plots of your first four transcripton factor inputs.
This tab displays information on where a TF is the most active and specific.
• The "By Cluster" sub-tab visualises the AUC of each TF in the selected cluster on the y-axis and the average AUC in all other clusters in the sample on the x-axis. User-selected fold change cutoff is displayed as a diagonal line and TFs with fold change greater than this cutoff is summarized in the table.
• The "By TF" sub-tab visualises the AUC of the first 4 selected TFs across all clusters in the sample. If there are more than 30 clusters in the sample, the 30 clusters with the highest AUC value for that TF is shown.