CytReg_v2 is a database of manually curated literature-derived interactions and empirically tested eY1H-derived interactions between transcription factors (TFs) and cytokine genes. CytReg_v2 contains 1,712 literature-derived interactions between 224 TFs and the regulatory regions of 116 cytokine genes, and 1,380 eY1H-derived interactions between 265 TFs and 108 cytokine promoters.
The CytReg_v2 website provides a user-friendly interface to explore the TF-cytokine interaction dataset by applying different user-selected filtering criteria. The user can select a set of TFs and/or a set of cytokines of interest using the standard HGNC nomenclature. Three types of basic searches that can be performed in CytReg_v2:
By default, searches will return interactions reported in any species by any type of evidence. Interactions can be filtered by human and/or mouse interactions using the “Species” check box. In addition, specific types of interaction evidence can be selected from functional assay (e.g. reporter assays), chromatin immunoprecipitation (ChIP), enhanced yeast one-hybrid assays (eY1H), and in vitro binding assay (e.g. EMSA).
In the case of predicted human interactions, SEEK (http://seek.princeton.edu/) was used to identify the top 100 co-expressed genes with the known cytokine targets of a given TF. Then, within the cytokines found among those 100 genes, the presence of a TF binding site was determined using motif analysis. These predicted interactions were classified according to confidence: high (two or more TF binding sites and evidence of interaction in the mouse cytokine GRN), medium (two or more TF binding sites but absent from the mouse cytokine GRN, or less than two binding sites but presence in the mouse cytokine GRN), and low (one binding site and absent from the mouse cytokine GRN).
TF expression patterns vary across cell-types, therefore filtering by TF expression thresholds is a useful approach to study TF-cytokine interactions in a more specific cellular context. Proteomics data is a valuable resource to quantify protein levels, however, few datasets are available for immune-related tissue types. In contrast, RNA-Seq is the most popular approach to quantify mRNA levels of genes across multiple cell-types. However, mRNA levels cannot be used to directly infer protein level as mRNA and protein abundance have not shown strong correlation in other studies. In order to facilitate the analysis of specific cell type cytokine regulation, CytReg_v2 incorporates external sources for mRNA and protein levels across multiple immune-related cell types:
CytReg_v2 displays results in a table format, containing the cytokine and TF interactions together with interaction-related information. TF and cytokine gene names are provided in HGNC format. The user can click a TF or cytokine to be directed to its UniProt page for further information. The type of interaction evidence is shown in the “Evidence” column based on the filtering criteria from the search page (see “Basic Search” section for evidence details).
TFs can activate and/or repress target gene expression. During the manual curation the regulatory activity of a TF impinging on the cytokine target was annotated (as activator, repressor or bifunctional). The “Species” field displays the species where the interaction was found (e.g., human and mouse) and also depends on the filter selected from the “Search” page. Finally, the PubMed IDs of papers reporting a given interaction are shown in the “Source” column. Users can click the IDs to be directed to the PubMed link of a given paper. These results table can be downloaded as a csv file using the “Export CSV Table” button.
The TF-cytokine interactions can be visualized as a network where the nodes represent TF and cytokines and the edges are the interactions. Node colors represent TF and cytokines and edge colors represent the type of interaction: activating, repressing, bifunctional, and physical (only evidence from binding but not functional assays). The complete list of colors for nodes and edges can be found in the network page. Note that edges are directional (TF to cytokine), as CytReg only reports TF regulating/binding cytokine regulatory regions. The interactive network is constructed using cytoscape.js, which allows moving nodes and zooming in and out for further customization.