KLIFS Frequently Asked Questions

• How can I cite KLIFS in a publication?
Currently there are two KLIFS articles, one in Nucleic Acids Research and one in ACS Journal of Medicinal Chemistry. Please cite both articles if possible.
When utilizing this website and this database or data derived from it we would like you to refer to the Nucleic Acids Research paper (database issue of 2016).
KLIFS was initially introduced in 2013 in a "perspective" in the ACS Journal of Medicinal Chemistry, please refer to this paper when discussing the analysis of kinase inhibitor binding mode classification, the binding site residue nomenclature, and the in-depth analysis and discussion regarding the large diversity of chemical scaffolds that are recognized by catalytic kinase domains.
• How can I use KLIFS?
A good place to start with KLIFS is via the quick-start tutorial.
• What software and data sources are being used to create KLIFS?
We are using in-house developed software (at the Division of Medicinal Chemistry, VU University Amsterdam) that combines many different software packages and multiple databases, but the heart of KLIFS has been built using MOE from the Chemical Computing Group analyzing structural data from the Protein Data Bank.
The interactive kinome is based on the KinomeRender developed and published by Chartier et al.. Kinome illustration (Manning et al.) reproduced courtesy of Cell Signaling Technology, Inc. (www.cellsignal.com).
• The 3D viewer does not seem to work.
The 3D-viewer (iview) is a WebGL viewer and requires no plugins to be installed. All up-to-date modern browsers support WebGL (more information), however, sometimes the configuration of the browser has to be adjusted. Please follow the steps described here to enable webGL for your specific browser.
• How did you define the binding site?
Based on our initial analysis of over 1200 kinase-ligand crystal structures we defined a pocket that comprises 85 residues and covers the interactions between inhibitors and the proteins of type-I, I½, II, III inhibitors and most type IV inhibitors as well. See the image below (and our publication) for more information.
• What do all the colors for the residues mean?
Each color represents a specific part of the kinase binding site (e.g. the hinge region is colored magenta, a-helices are colored red, B-sheets are colored yellow, the gatekeeper is colored gold). See the image below (and our publication) for more information.
• What are these different (sub)pockets?
The binding site of the catalytic domain of kinases can be roughly divided into three main pockets: the front pocket, the gate area and the back pocket. Several kinases also have specific subpockets that have been classified into front pockets I and II, back pockets I-V. A schematic representation is given below (see our publication for more information):
• Conserved waters or pocket waters: what do these names like I1, O2 mean?
Based on the overlay of all kinase structure we have analyzed the presence of (highly) conserved waters that frequently interact with the co-crystallized ligand. This resulted in the identification of 13 positions in which water appear in kinase structures with a high frequency. These positions have been named I1-I11 and the DFG-out specific structures have been classified as O1 and O2.The position of these water clusters are highlighted in the figure below:
• What is an MOE database?
MOE stands for Molecular Operating Environment and is a fully integrated drug discovery software package of the Chemical Computing Group. The MOE database download processes all the selected structures into a single fully annotated database file for this software package. See their website for more information: Chemical Computing Group
• What are these alternates/alternate models?
Alternates are alternative conformations or rotamers that can be defined for a single residue/substance/ligand in a crystal structure. A crystallographer can define alternates when the electron density for these elements are not concentrated on a single spot, but are clearly present on 2 or more spots. Each of the models will then be stored in an alternate (A, B, etc.). However, most PDB-viewer only load the first (A) model discards the rest of this information. As this information can be valuable we have chosen to split the alternates and store them in separate mol2 files.
A great example of two alternates for a ligand and residues is the X-ray of MAP kinase p38 in complex with a pyridinylimidazole (PDB-code 2EWA). In the figure below this structure is depicted for both alternates in exactly the same orientation:
• How can I download (all) processed structures from KLIFS?
It is possible to download the processed kinase structures in MOL2 format or as a MOE database after performing a search using the download button at the bottom of the search results page. If you do not provide any search/filtering criteria KLIFS will return all kinase entries within the database, therefore allowing you to download all structures from KLIFS. All related annotations (IFP, pocket alignment, resolution, pockets, water clusters, etc.) are embedded within the MOE database or in a CSV file (overview.csv) within the MOL2 package.
• Where can I download the master sequence alignment?
The master sequence alignment has been optimized for the correct alignment of the 85 pocket residues. This alignment is therefore not suitable to be applied in a different setting. However, for a good overall alignment of kinases, we recommend using the kinase alignment of Manning et al., which is freely available at kinase.com.
• I cannot find a specific crystal structure in KLIFS that is present in the PDB, what happened and can you add it?
Please make sure that the PDB structure contains a human or mouse kinase (these are the major species with crystallized kinase domains, others will follow in the future). Please send us the information regarding this structure and we will make sure to add the structure and check why it was not present in KLIFS already (click on the contact button at the bottom).
• Can I combine different search options with each other?
Yes, all search options that are shown can be combined. So it is possible to search for an inhibitor that, for example, shares similarity with Dasatinib, belongs to the Ack family, targets the backpocket I B, contains a conserved water at cluster i5, contains a Leucine at the GK+2 position (residue 47) and has an H-bond interaction with the gatekeeper with a minimal crystal structure resolution of 2.5.
• I have identified an error (e.g. in the annotation, alignment, atom typing of a structure), can you correct it?
Please send us all information regarding the error using the Give us feedback button at the bottom of the webpage for this specific structure and we will correct it as soon as possible (or you can use the contact button at the bottom of this page).
• Can you please add a specific analysis or feature to your database/website?
Please contact us with your ideas for the improvement of KLIFS (click on the contact button at the bottom).
• Contact us!
You can contact us via this form: