User guide
Search, explore and download protein–protein interaction data.
Searching and exploring networks¶
Click a section to expand it.
Starting a search¶
What to type, phrases, suggestions and example searches.
Type one or more proteins into the search box on the home page, or the Query box beside the results, then press Search or Enter.
- What to type: gene names such as
TP53work on every portal. Other identifiers, such as Ensembl gene IDs, work where the portal's data records them. Capitalisation does not matter, but a term must match a whole identifier:TP5finds nothing. - Several proteins: separate them with commas, spaces or new lines, for
example
BAD, BCL2L1, BAK1. Shift+Enter adds a new line instead of searching. - Suggestions appear after two characters. Use ↑/↓ and Enter to pick one, or Esc to close the list.
- Example searches below the box run a ready-made search chosen by the portal's administrator.
The address of the results page is …/search/<your terms>, so you can
bookmark a search or send it to someone.
Typing a phrase¶
The home page box also understands short phrases, such as:
BCL2 in liverTP53 and MDM2 with high confidencewhat binds CDK2 in testis
From a phrase, openPIP picks out the proteins, any tissues you name (which become the tissue filter), and confidence wording. High confidence, confident or reliable mean a score of at least 0.5. Score above 0.7 sets that exact threshold. A line under the box shows what will be searched and which filters will apply. Words it recognises but cannot act on, such as two-hybrid or bait, are listed after "Cannot filter by". Only the home page box reads phrases. The Query box beside the results takes a plain list of proteins.
Found and not found¶
Beside the results, Found lists the terms that matched and Not found the ones that did not. If nothing matched, the network area says so. Check the spelling, or try the protein's gene name.
The results page¶
Filters, the network and the result tables.
The page has three parts: the controls, the network, and the result tables underneath. The administrator chooses whether the controls sit in a column on the left or in a row of headings above the network. In the row layout, each heading opens its panel when you point at it or click it.
Drag the bar between the network and the tables to resize them. Click the bar to reset it, or use ▲/▼ to make the network short or tall. Fullscreen, at the top left of the network, gives it the whole screen.
Filters¶
Filters change what the network, the tables and your downloads show.
| Filter | What it does |
|---|---|
| Min. confidence score | Hides interactions scored below the slider (0–1). Interactions with no score count as 0, so they disappear as soon as the slider is above 0. |
| Interaction sources | One checkbox per interaction category in the results, such as Published or Literature. Untick a category to hide its interactions. |
| Filter mode | No Filter shows every interaction among the proteins found, including between partners. Query-Query keeps only interactions between proteins you searched for. Query-Interactor keeps interactions with at least one protein you searched for. |
| Tissue expression | Keeps only proteins expressed in every tissue you tick. Only tissues with expression data for these results are listed. All tissues clears the selection. |
Two switches in the tissue panel change how proteins are drawn: Show tissue expression sizes each protein by its expression level, and Reflect tissue specificity shades it by how specific that expression is: lighter for low specificity, darker for high, in shades of the protein's usual colour. Interaction colours do not change. While either switch is on, you can select only one tissue.
Filters stay set when you start a new search.
The network¶
- Proteins you searched for are drawn in the query colour, and their partners in the interactor colour. Point at a protein to see its name.
- Interactions are coloured by their highest interaction category. The legend at the top right shows what each colour means.
- Layout (bottom right) offers Force-directed (Cola), the default, Force-directed (CoSE), Concentric, Circle and Grid. The same menu sets the network's background colour for you alone. Your browser remembers it.
- Scroll to zoom, and drag the background to move around.
Click a protein to open its panel:
- 3D Structure shows the AlphaFold model, coloured by model confidence, or an experimental structure from the PDB when one exists.
- Actions: Search … for
starts a new search for that protein alone. Remove From Network hides it and its interactions from this view, from Save Network and from your downloads. The tables below still list it. - Links go to NCBI Gene, the Human Protein Atlas, Ensembl, GeneCards and UniProt. You also see the protein's interaction counts and description.
Click an interaction to see its confidence score, the datasets it comes from (with links to their publications), experimental details, and the literature that reports it.
Open in Cytoscape (the small network icon, bottom right) sends the network to Cytoscape running on your own computer. Cytoscape must be open, with its REST interface on the default port.
Result tables¶
| Tab | Contents |
|---|---|
| Interactions | One row per interaction: both proteins (linked to NCBI Gene), score, category and datasets. Click a column heading to sort. 25 rows per page. |
| Interactors | One row per protein: gene name, UniProt and Ensembl IDs, whether you searched for it, and its interaction count. |
| Molecular Function, Biological Process, Cellular Component, Reactome, CORUM, KEGG | Enrichment of the proteins in the network, calculated by g:Profiler with a false-discovery-rate threshold of 0.05. The top 50 terms are shown. |
| Subcellular Location | Where the proteins in the network are found in the cell, from the Human Protein Atlas. |
| Tissue Expression | Which proteins in the network are expressed in each tissue, from GTEx. |
| Protein Info | Everything about the protein you last clicked: identifiers, interaction counts, description, annotations, top tissues, sequence (with Copy FASTA) and 3D structure. |
Click a row in an enrichment, location or tissue table to highlight its proteins in the network. Click it again to clear the highlight.
The enrichment tables are calculated by g:Profiler, an outside service, so they need an internet connection. A portal can turn off the location and tissue tabs if its organism has no such data.
External tools¶
Send the network's proteins to g:Profiler, STRING, Reactome and more.
External links sends the proteins in the network to g:Profiler, Reactome, GeneMANIA, Pathway Commons, DAVID, STRING, cBioPortal, Complex Portal, Drugst.One or Genelist. Links to IntAct, BioGRID, KEGG and UniProt use only the proteins you searched for.
Keeping and sharing a search¶
Where saving, sharing and downloading are covered.
See Accounts, saving and sharing. Downloads are covered in Downloading data.