Frequently Asked Questions
|
PROMISCUOUS is a database developed to support scientists in drug-repositioning. It contains three different types of entities: drugs,
proteins and side-effects. These entities are connected between each other through drug-protein, protein-protein and drug-side-effect relations.
Proteins are retrieved from UniProt and are displayed with synonyms and organism information. Additional information as 3D-structures from PDB and EC-numbers are given if available. Drugs gained from SuperDrug were integrated to PROMISCUOUS and assigned metadata as ATC-codes (Anatomical Therapeutic Chemical), structure information and synonyms. The side-effects contained in the database were fetched from SIDER and related to the drugs. |
|
The aim of PROMISCUOUS is to offer a comprehensive datasets. Here for drug-target relations from different well-known databases
as DrugBank, SuperTarget and SuperCyp were integrated. To enhance the completeness of the dataset additional newly explored
relations were incorporated, too. The new information was obtained in two steps: Firstly, text-mining algorithms were applied
to sort all PubMed listed papers by their relevance for drug-target relations. In a second step, the 7,000 papers with the
highest rank were manually revised.
Protein-protein interaction data was obtained from ConsensusPathDB which integrates physical protein-protein interactions, metabolic and signaling reactions and gene regulatory interactions. Information on complex composition comes from Corum a protein complex database. |
|
To put the drug target relations contained in the database into a cellular context, the information was mapped onto KEGG Pathway maps. Targets for which drug information is available are marked in yellow. Targets are displayed in different colors depending on their characteristics as shown below.
To get additional information about the targeting drugs place the mouse pointer on the target. Drugs are shown with their name and structure. An additional table summarizing the Drug-Target relations is shown at the bottom. Like this the user can retrieve information on drug target relations in a pathway and organism specific way. The knowledge about on which pathways a drug acts can be very helpful in drug repositioning. |
|
For a scientific but yet intuitive way of exploring and handling the data set a novel interface was developed that offers a “natural” way of exploring a network.
The interface represents database entities (drugs, targets and side effects) as nodes in a network with edges representing the relations between them. By exploring the data set the user creates a unique drug-target-side effects network.
By double clicking a node neighbor node data is loaded into the network in real time. That allows users to construct “their own” complex networks of drug-target-side effects data. Detailed information of a node can be
requested by highlighting a node and using the button “open details in new tab”.
|
|
A spring logic is included in the network interface to ensure that the edges and nodes within the network organize in the clearest possible way. The spring constant can be altered to relax or aggregate the network and the network can be fixed if
a manual layout is desired (4). If the network becomes too big or confused the user can delete nodes from the network. Note that deleting nodes from the network causes neighbors of the deleted node to float away. To circumvent this effect the user can simply hide and unhide nodes (5).
|
|
|
During every step of constructing a user defined network it is possible to safe the current view as a XML file to the hard disk and to reload this XML into the interface again (2). This makes PROMISCUOUS not only a tool for viewing the
data set but makes it a tool to enable scientists to gain most possible information out of the data set, save the current result of work, discuss the current work and to alter and enhance it at later time.
|
|
The interface is divided into two parts. The main window shows the network representation of the currently loaded data set (1). The right panel (3) gives context dependent detail information, links to detail information of highlighted nodes
or lists of related entities (e.g. drugs indicated for the same highlighted side-effect). This way of representing the data shows in our opinion a completely new, user friendly yet exhaustive way that not only enables one to view
the data but also to give insights into the complex interaction behavior of the given entities and therefore also gives a starting point for more complex analysis as the identification of yet unknown drug-target-side effects relations (drug repositioning).
For first insights of graph theoretical analysis basic average graph properties (Betweensess, Degree, Clustering Coefficient) for the current and the overall network can be calculated.
|
|
The fulltext search is our general search interface. It makes it possible to have a quick start into PROMISCUOUS and to search for drugs,
targets and interactions at once.
You have the option to decide, whether all query results or only those with known interactions are displayed. An additional input
box for the fulltext search is also displayed at the bottom of every other site of PROMISCUOUS.
Results are displayed ordered into the categories target, drug and interaction. Found objects are linked to the related detail sites.
|
|
Drugs can be searched by different identifiers. The query results are shown in a table with their name and PubChem id. To get additional information on the drug, click on its name, you will be redirected to a drug detail site. Synonyms, ATC-Codes and Targets are available on-click as well. To see the chosen drug in the network visualization, click on the network icon  . To see in which pathways the targets addressed by the drug are involved click on the pathway icon  . Drugs can be searched by: - Drug name - PubChem Id - ATC-Code - Side Effects It is possible to choose whether only drugs with known targets or all drugs are shown by putting a checkmark on this option. |
|
Targets can be searched by different identifiers. The query results are shown in a table with their name, Uniprot ID, Uniprot Accession number, organism and if available EC-number. To get additional information on the target, click on its name, you will be redirected to a target detail site. Synonyms and drugs acting on the target are available on-click as well. To see the chosen target in the network visualization, click on the network icon . To see in which pathways the targets is involved click on the pathway icon . Targets can be searched by: - Target name - EC-number - Uniprot ID - Uniprot Accession number - PDB ID - Kegg ID To choose whether only targets with known drugs or all targets are shown, the user has to put a checkmark on this option. |
|
The pinboard is a search help to facilitate the exploration of the data. It has two main functions.
In first instance it is possible to save performed queries in the pinboard, to repeat them easily further on.
Targets, drugs, interactions, or pathways can be added to the pinboard. The other important function of the pinboard
is to load several drugs or targets into the network visualization at once. Every result in PROMISCUOUS is displayed with a pinboard button  .
By clicking on the icon the object is added to the pinboard and can be found under Pinboard. It is displayed with its
name and category.
To get back to the detail view of the pinned entry click on its name. The objects can be removed from the pinboard
separately by clicking on the bin icon  or all at once by clicking on the clear button.
In order to load more than one element into the network visualization a check mark has to be set befor clicking onto
the "show selected entities in the network" button. |
|
To be able to use the applet, you must have Java properly installed and your Browser must have the permission to show
Java-applets. There is a bug in the new Java version 7. If you open the Java options dialog and remove the check mark saying that Java can save temporary data locally the problem should be solved. If you still have difficulties with the Applet you might be using openJDK as your Java runtime environment. We experiences problems with using openJDK under linux on some computers. The best solution to this problem is to use Sun java as your runtime environment. We did not have any problems while testing the page under MacOS, Windows and Linux using Sun's java. If you are experiencing problems with one of these configurations, please do not hesitate to mail us about it. |
We were not able to map all drugnames to specific PubChem compound Ids as multiple compounds have similar names. In these cases, the
known interactions for one drug were mapped to all compounds with the same name. The list of names where this applies is shown below.
|
|
The data stored in PROMISCUOUS forms a valuable resource to analyse the relations between drugs, targets and side-effects.
Currently the data is available through the website. If this is not sufficient for you to perform your scientific work, please
do not hesitate to contact us.
|
This work is licensed under a Creative Commons Attribution-Noncommercial-Share Alike 3.0 Unported License.
