To go back to the TissueNet v.2 home page click here.
This tutorial offers an overview of TissueNet v.2, explanation of the icons which are constant across TisseuNet v.2 pages, explanation about the different search options and the TissueNet v.2 output page.
Please see our FAQ page.
This website is best viewed at a resolution of 1440×900 and above.
Contact: Esti Yeger-Lotem at estiyl@bgu.ac.il
Table of contents:
The TissueNet database (https://netbio.bgu.ac.il/tissuenet/) provides quantitative tissue-associations for human protein-protein interactions (PPIs). These associations are important for identifying tissue-specific roles of human proteins and for shedding light on tissue processes and phenotypes. However, many experimentally identified PPIs have no tissue-context. To associate PPIs with tissues, TissueNet harnesses extensive expression profiles that were measured via RNA-sequencing and protein-based assays from tens of human tissues. Users can select a protein, a tissue, and an expression data source, and obtain a network view of the query protein and its tissue-associated PPIs. The current version of TissueNet gives users the flexibility to select expression data sources and to set interactively the expression threshold for tissue-association, such that interacting proteins whose expression is below the threshold are ignored. The output network highlights proteins that are tissue-specific and globally-expressed according to the threshold set by the user, and thus enables users to quickly assess the tissue-specificity or global functionality of the network. The user can also switch to a quantitative view that highlights proteins that were significantly up- or down- regulated in the selected tissue relative to all other tissues. Thus, TissueNet 2 offers an extensive, user-friendly and quantitative interface to study the tissue-specificity of human proteins and their interactions.
Icons Appearing on TissueNet v.2 web-pages:
These icons are constant across all TissueNet v.2 pages:
– to TissueNet v.2 homepage from every page.
– open TissueNet v.2 homepage in a new tab.
– to tutorial from every page.
PPI data were gathered and from four major databases (BioGrid, DIP, MINT and IntAct) using the MyProteinNet webserver [1], currently resulting in 243,706 interactions between 17,283 proteins.
Note: Most of the available PPI data are oblivious to alternatively-spliced isoforms. Therefore, TissueNet 2 associates each protein-coding gene with a single protein product. For simplicity we refer to the gene and its protein product interchangeably.
- Show the tissue-associated protein interactions for your protein: The user selects an expression dataset, tissue, protein and expression threshold. The output is a network view of the query protein and its PPIs in the selected tissue. Only proteins whose expression level is ≥ the threshold will be shown.
- Show the tissue-associations for your protein-protein interaction: The user selects an expression dataset, tissue, a threshold, and two interacting proteins. The output shows this interaction in the selected tissue only if both proteins are expressed at a level that is ≥ the threshold, otherwise a warning is given.
TissueNet v.2 uses tissue expression profiles from three datasets:
- The Genotype-Tissue Expression (GTEx) consortium [2]. 421 RNA-sequencing profiles from 42 tissues were included. These samples were taken from individuals with death reason 1 (traumatic injury), and not from treated patients.
- The Human Protein Atlas (HPA) project [3]. 192 RNA-sequencing profiles from 27 tissues were gathered. In both datasets, raw counts were normalized for each data source using the TMM method by the edgeR package [4], to obtain the same library size for every sample. Genes with less than 1- raw counts in all samples were removed before normalization. For each tissue the normalized count for a gene was set to its median normalized count across all corresponding tissue samples.
- HPA protein expression dataset [3]. Included expression levels for 14,578 proteins in 83 tissues, which were united into 47 merged tissues (e.g., appendix glandular cells and appendix lymphoid tissue were merged to Appendix). The expression level of a protein in merged tissue was set to its highest level in the relevant samples.
Once a dataset is selected, the user can select a specific tissue from this dataset to obtain its associated PPIs. The tissues drop down list is searchable for convenience.
The output contains the query proteins and its PPI partners whose expression level was at least as high as the selected expression threshold in the selected tissues. If the user selected ‘All Tissues’ a protein has to be at least as high as the threshold in any tissue.
The user can search for a protein by HUGO Gene Symbol, Ensemble gene IDs or by Entrez gene IDs. After the user has entered the gene name into the text box TissueNet will query its server check if the gene is in the database. If the gene is not in the TissueNet database, the message “Your search returned no results” will be displayed. If the gene was found in the TissueNet database, it should be selected from the displayed list.
TissueNet 2 offers the user the option to set the expression threshold dynamically. Only proteins whose expression is ≥ threshold will be presented in the network. Proteins with expression level lower than the threshold will be removed from the network along with their interactions.
For RNA-sequencing datasets (GTEx and HPA), expression is measured by normalized counts per million (CPM). In the protein expression dataset, expression level ranges from no expression (0), low (1), medium (2) and high (3) expression. Selecting a threshold of 0 will show all the known interactions of the protein.
This link displays the DMD protein in GTEx dataset with “All Tissues” selected at a threshold of 8.0 normalized CPMs.
This link displays a randomly selected protein. The expression dataset, tissue and threshold are all set randomly.
TissueNet 2 output page offers several parts: The Network View, the Tabs View and the Mouse right-click menu. Here we offer explanations to these parts.
The network view is an interactive graphical representation of the output, displayed using Cytoscape.js plugin [5]. Nodes represent proteins, and edges between them represent their experimentally detected tissue-associated PPIs.
The first output is always displayed in the tissue-specificity view. In this view, proteins are colored to highlight tissue-specific proteins, expressed in at most 20% of the tissues, and globally expressed proteins, expressed in at least 80% of the tissues. This view enables users to quickly assess the tissue-specificity or global functionality of the output network.
Since GTEx contains samples from tissues with multiple regions (e.g., 11 brain sub-regions) versus tissues with one region (e.g. Liver, Lung), only to compute tissue-specificity the different sub-regions were counted as one (e.g., brain sub-regions were counted as one brain tissue).
Figure 1. Tissue-specific view legend
Table 1. Tissue-specificity coloring values
| GTEx – RNA-Seq | HPA – RNA-Seq | HPA – Protein Expression | |
| Tissue Specific | 1-4 Tissues | 1-5 Tissues | 1-9 Tissues |
| Intermediate expression | 5-17 Tissues | 6-21 Tissues | 10-35 Tissues |
| Globally Expressed | 18-22 Tissues | 22-27 Tissues | 36-47 Tissues |
The quantitative view is obtained by selecting it in the interactive menu. It highlights proteins that were significantly up- or down- regulated in the selected tissue relative to all other tissues. This view is only available for the GTEx and HPA RNA-sequencing datasets, for which enough samples were available. It is not available when ‘all tissues’ are selected.
Figure 2. Quantitative view legend
i: The information button
Users can retrieve various attributes of selected proteins and interactions through the tabs view on the right. The available tabs are Properties, Gene Ontology, Tissue Expression and Graph Options.
This tab shows information about selected proteins and interactions.
Proteins have the following attributes:
- HUGO Gene symbol of the protein.
- Ensembl Gene ID of the protein.
- Entrez ID of the protein.
- # Main Tissues where the protein is expressed. Tissues are group based on main tissues to prevent a bias towards tissues in which many sub tissues where sampled as explained in the network view section.
- Expression Value in the selected tissue, in normalized counts.
- Log2 Fold Change of the selected gene in the selected tissue relative to all other tissues.
- P-value of the selected gene in the selected tissue relative to all other tissues.
Interactions have the following information:
- Protein 1: HUGO gene symbol of one of the interacting proteins.
- Protein 2: HUGO gene symbol of the second interacting protein.
- # Main Tissues associated with this interaction.
- Detection Method(s): Provides the experimental method(s) by which the interaction was detected (e.g., Affinity Capture-Western); whether this was a low-througthput (Low) or high-throughput (High) experiment; and the database reporting this PPI (e.g. BioGrid).
This tab shows Gene Ontology (GO) annotations for selected proteins. Each GO annotation appears with its GO id, description and evidence code. GO annotations are retrieved from AMIGO using the MyGene.info web-service [6].
This tab provides the expression values for selected proteins in different tissues. Tissues in which the protein is expressed below the threshold are highlighted in red.
For selected interactions, this tab shows only tissues in which both pair-mates are expressed above the selected threshold.
This tab offers several ways to manipulate the network view.
Change Network Layout: allows the user to select a different network layout. The supported layouts are: (i) spread, in which the proteins are spread evenly around the query protein, (ii) random, in which proteins appear in random order in the network view, (iii) grid, in which proteins are ordered evenly on a grid, (iv) circle, in which proteins are ordered on a circle and (v) concentric, in which the query protein appears in the middle and the rest of the proteins are ordered on a circle around it.
Network Options: Includes three features
(i) Remove Selected Proteins – remove the selected proteins and their interactions from the network
(ii) Select an additional query protein – upon selection of a network protein, adds its tissue-associated PPIs to the current network. This option is available for a single protein at a time,
(iii) Export to GraphML – download a GraphML file for the network, to be used with other network applications, such as Cytoscape.
Filter network by protein expression across tissues: filter tissue-specific proteins (expressed in ≤ 20% of the main tissues); ubiquitous proteins (expressed in ≥ 80% of the main tissues); and other proteins.
Clicking the right mouse button opens a menu with the following options:
1. Select as an additional query protein: Upon selecting a protein in the network, this option allows you to add its tissue-associated PPIs to the current network. (this option allows the addition of a single protein each time).
2. Remove Selected Proteins removes your selected proteins and their PPIs from the network.
3. Export to GraphML lets you export the information of the proteins and PPIs to GraphML format. This format can be loaded into Cytoscape.
4. Change Layout lets you choose a different layout for the graph. By default, spread layout is selected.
Improvements over the previous version of TissueNet:
The current version features significant data expansion and enhanced usability. Firstly, the number of experimentally detected PPIs available in public databases has tripled since TissueNet was published in 2013. Secondly, the coverage and quality of RNA-Sequencing data increased considerably. We replaced the TissueNet gene expression datasets that supported association to 16 main tissues, with RNA-sequencing data from the Genotype Tissue Expression (GTEx) consortium [2] (421 samples from 42 tissues), and the Human Protein Atlas [3] (192 samples from 27 tissues [3]). Thirdly, protein expression screens (HPA, [3]) expanded to include more than 14,000 proteins measured in 47 tissues. Fourthly, the increase in RNA-sequencing data provides a rich quantitative view, by selecting this view the output network is colored by the expression fold-change of the genes in the selected tissue. Lastly, while TissueNet applied a fixed threshold for pair-mates co-expression, TissueNet v.2 supports dataset-specific, user-selected thresholds for associating PPIs to 27-47 main tissues. To conclude, TissueNet v.2 offers significant enhancements over TissueNet in terms of data quality, coverage and user flexibility.
REFERENCES:
1. Basha, O., et al., MyProteinNet: build up-to-date protein interaction networks for organisms, tissues and user-defined contexts. Nucleic Acids Res, 2015. 43(W1): p. W258-63.
2. Consortium, G., The Genotype-Tissue Expression (GTEx) pilot analysis: multitissue gene regulation in humans. Science, 2015. 348(6235): p. 648-60.
3. Uhlen, M., et al., Proteomics. Tissue-based map of the human proteome. Science, 2015. 347(6220): p. 1260419.
4. Robinson, M.D., D.J. McCarthy, and G.K. Smyth, edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics, 2010. 26(1): p. 139-40.
5. Franz, M., et al., Cytoscape.js: a graph theory library for visualisation and analysis. Bioinformatics, 2016. 32(2): p. 309-11.
6. Wu, C., I. MacLeod, and A.I. Su, BioGPS and MyGene. info: organizing online, gene-centric information. Nucleic acids research, 2013. 41(D1): p. D561-D565.
7. Kuhn, M., et al., STITCH 3: zooming in on protein-chemical interactions. Nucleic Acids Res, 2012. 40(Database issue): p. D876-80.