Database access and features
The HTRIdb is freely accessible via a user-friendly web interface at http://www.lbbc.ibb.unesp.br/htri. Besides searching for TF-TG interactions data of interest (see below), users can (i) download all interactions–TF-TG or TF-TG with PPIs–contained in the HTRIdb through the “Download” page, (ii) keep track of the number of TFs, TGs, TF-TG interactions, articles and techniques present in HTRIdb and verify the proportion of known human TFs covered by HTRIdb through the “Statistics” page, (iii) learn how to retrieve and download or visualize as a network the TF-TG interactions and PPIs of a given TF or TG of interest via the “Tutorial” page, (iv) add new TF-TG interaction data to HTRIdb, as previously mentioned, by uploading their newly discovered TF-TG interactions data via the “Upload Data” page, and (v) send suggestions and comments or point out inconsistencies encountered in some TF-TG interaction to HTRIdb staff via the “Contact Us” page (Figure ).
Regarding the search mechanism (for a simple use case, please see the next section), users can retrieve TF-TG interactions data on a specific TF or TG of interest. For this purpose, users should access the “SEARCH” page and then proceed to the “TRANSCRIPTION FACTOR” or to the “TARGET GENE” page, depending on the type of search (Figure ). In these pages, queries can be entered using the Entrez Gene ID code or the total or partial TF or TG gene official or alias symbols (Figure ). This feature eliminates the need for users to know the exact name of TF or TG of interest; moreover, a page containing a list of possible TF or TG gene official symbols matching the entered query further helps users to confirm their choice (Figure ). The retrieved data includes the Entrez Gene IDs, the gene official symbols and the links to the respective Entrez Gene report pages for both TFs and TGs and the name of experimental technique used to detect the interactions with the Pubmed IDs for articles reporting the discovery of these interactions (Figure ). TF-TG interactions data only or TF-TG interactions data associated with their PPIs can be downloaded in a tab-delimited or spreadsheet file; moreover, TF-TG interactions can be visualized as a network by clicking the “Graph” icon (Figure ).
Figure 3 Screenshot of the “Search” page. In this page, users should click the “TRANSCRIPTION FACTOR” or the “TARGET GENE” buttons if they want to search TF-TG interactions data for, respectively, a TF or a TG of (more ...)
Figure 4 Screenshot of the “Transcription Factor” page within “Search” page. In this page, users should select the type of search key. If they know the TF’s EntrezGene ID, the “NCBI GeneID” option should (more ...)
Figure 5 Screenshot of the page to select and confirm the TF of interest. This page shows a list of possible TF’s gene official symbols matching the entered query in the “Transcription Factor” page (in this case, “PAX”–see (more ...)
Figure 6 Screenshot of the result page. The result page displays all TF-TG interactions related to the TF of interest (in this case, PAX8). It contains TF’s NCBI GeneID, TF’s official gene symbol, link for additional information on TF in the Entrez (more ...)
In the network visualization tool provided by the embedded Cytoscape Web software [8
], networks can be seen in three levels (Figure ). The first level displays only the interactions between the TF of interest and its TGs (Figure a), the second level displays interactions from first level plus PPIs between TF of interest and other proteins (Figure b) and the third level displays interaction from first and second levels plus PPIs between TGs and other proteins (Figure c). The network displayed in the second and third levels can be downloaded in text format by clicking the“Download Network” icon (Figures b and c).
Figure 7 Screenshots of the network visualization pages. This tool is provided by the embedded Cytoscape Web software  and shows networks in three levels: first level displays only the interactions between the TF of interest and its TGs (A), second level displays (more ...)
Through the network visualization of TF-TG interactions, users can also check the reliability of interactions. Reliability scores were assigned to interactions according to the number of articles reporting the interactions and the number of different techniques used to detect them (Table ). These scores are coded with a color gradient from light blue (Reliability 1) to dark blue (Reliability 4) (Figure ).
Reliability scores for TF-TG interactions
Case study: searching for PAX8 target genes
In this section, we describe the steps for retrieving the target genes of TF PAX8 and then visualizing the interactions between this TF and its TGs along with associated PPIs as a network. This case study serves both to reinforce key features of HTRIdb and as an example that could be followed to interrogate any TF or TG of interest.
1. In the main page of HTRIdb, click the “SEARCH” button located in the left menu (Figure );
2. In the “SEARCH” page, select the“Transcription Factor” button (Figure );
3. In the “TRANSCRIPTION FACTOR” page, select the type of search key. If you know the TF’s EntrezGene ID, select the “NCBI GeneID” option and enter it in the search field; otherwise, select the “Gene Symbol” option and enter the complete or partial TF’s official gene or alias symbol. In this case, the TF of interest is PAX8 and the user enters the TF’s gene partial symbol “PAX” (Figure );
4. After clicking the “Search” button in the “TRANSCRIPTION FACTOR” page, a page containing the list of possible gene official symbols matching “PAX” appears (Figure ). In this page, select the official gene symbol that represents your TF of interest (in this case, PAX8) and then click the “Search” button. By clicking “[+]”, it is possible to see the selected TF’s alias symbols (Figure ).
5. The result page displays all TF-TG interactions related to PAX8 (Figure ). To download TF-TG interactions data only or TF-TG interactions data associated with their PPIs, select the interactome type in the selection box in the “Download” area and then click the icons below the selection box to download the data in text format or in spreadsheet format (Figure );
6. To visualize the PAX8 and its TGs along with associated PPIs as a network, click the “Graph” icon in the result page (Figure ) and then navigate through the levels of the network visualization tool (Figure ). The first level displays only the interactions between PAX8 and its TGs (Figure a), the second level displays interactions from first level plus PPIs between PAX8 and other proteins (Figure b ) and third level displays interaction from first and second levels plus PPIs between the PAX8 TGs and other proteins (Figure c).
Comparison to other related databases
As mentioned in the “Introduction”, the most prominent examples of TF-TG interactions databases are JASPAR [2
], the Open Regulatory Annotation database (ORegAnno; [3
]), Swissregulon [4
], TRANSFAC database [5
], the Transcriptional Regulatory Element Database (TRED; [6
]) and the Transcription Regulatory Regions Database (TRDD; [7
]). As JASPAR and Swissregulon are mainly focused on TF-DNA binding sites, we will compare our database only with those databases from which experimentally verified TF-TG interactions can be extracted, namely TRANSFAC, TRRD, TRED and ORegAnno.
Despite their remarkable usefulness as sources of experimentally verified human TF-TG interactions, TRANSFAC, TRED, TRRD and ORegAnno present some constraints that limit their use to construct human transcriptional regulatory networks in a feasible way in comparison to HTRIdb. We discuss below the advantages of HTRIdb over these databases.
Although TRANSFAC is considered the leading TF-TG interactions database, the advantage of HTRIdb over TRANSFAC is that HTRIdb is freely accessible. To take advantage of all HTRIdb features, including the downloading of all data present in HTRIdb, users are not required to make any subscription. On the other hand, as TRANSFAC is marketed as a commercial resource, users are required to make a paid subscription to access its contents and, accordingly, to download its TF-TG interactions.
TRED and TRRD are freely accessible as HTRIdb is, but their major drawback is that they do not provide links to download their collection of TF-TG interactions. Conversely, HTRIdb, besides providing links to download the displayed results (Figure ), also provides links to download all TF-TG interactions through the “DOWNLOAD” page that is easily accessible from the main page of HTRIdb (Figure ). Besides the presence of download links, HTRIdb also offers additional advantages over TRED and TRRD, such as a dynamic network visualization tool, which is absent in TRRD and TRED (in fact, TRED provides only static images showing the interactions), and a more friendly result page. While the result page of TRED is broken into subpages containing a maximum of 20 TGs each in which techniques and Pubmed IDs are not displayed (Figure b), the result page of HTRIdb displays all TGs of the TF of interest along with the techniques used to detect the interactions and the Pubmed IDs for articles reporting the interactions (Figure and Figure a). Regarding the result page of TRRD, links to pages showing further information on TF-TG interactions of interest are all displayed at once, but the names of links are non-obvious internal identifiers used by TRRD.
Figure 8 Screenshots of the result pages of HTRIdb, TRED and ORegAnno. The result page of HTRIdb displays all TF-TG interactions data in one page only (A), while the result pages of TRED (B) and ORegAnno (C) show partial TF-TG interactions data and are broken (more ...)
ORegAnno is also freely accessible as HTRIdb and TRED are and, differently from TRED and similarly to HTRIdb, provides links to download its entire collection of TF-TG interactions. The advantages of HTRIdb over ORegAnno mainly concern the following features: the format and quality of data contained in the downloaded file and the layout of the result page. With regard to format and data quality, ORegAnno tab-delimited flat files should be heavily parsed to remove non-human interactions, entries with missing information and gene name ambiguity; regarding the ORegAnno result page, it is broken into subpages containing a maximum of 10 TF-TG interactions each in which techniques used to detected the interactions are not displayed (Figure c). To find the technique, users should click the “RECORD DETAILS” link (Figure c). As already discussed above, the result page of HTRIdb displays all TGs of the TF of interest along with the techniques used to detect the interactions and the Pubmed IDs for articles reporting the interactions (Figure and Figure a). In addition to the format and data quality of the downloaded flat file and layout of result page, HTRIdb also presents other two advantages over ORegAnno: (1) the network visualization tool and (2) a higher number of human TF and TG entries. While HTRIdb has 284 TFs and 18302 TGs, ORegAnno has 134 TFs and ~ 1800 TGs.