GLASS (GPCR-Ligand Association) database is a manually curated repository for experimental data pertaining to GPCR-ligand interactions. Along with relevant GPCR and chemical information, GPCR-ligand association data are extracted and integrated into GLASS from literature and public databases.
A list of currently-known GPCRs was compiled from UniProt and used to filter through the other chemical databases for ligand-association data (ChEMBL, BindingDB, IUPHAR, DrugBank, PDSP), GPCR diseases association (TTD), GPCR experimental structural data (PDB, BioLiP), and predicted models of GPCRs (GPCRRD). Subsequently, information from the extracted databases were unified to the same format and checked to ensure that all entries are only GPCR-related. Thus, the user would not find any entries on receptor tyrosine kinases or any other protein that is not a GPCR. All relevant ligand chemical data (PubChem) and GPCR data (UniProt) were extracted accordingly for each GPCR-ligand entry. Each molecule with a unique InChI key was considered a unique ligand entry in the database.
In GLASS, you can:
XlogP is a measure of the molecule's lipophilicity (molecules with high XlogP are lipophilic and will reside in the cell membrane) and solubility (molecules with high XlogP tend to be insoluble). XlogP is determined by the predicted partitioning of the molecule into hydrophobic octanol in a mixed octanol/water system.
Druglikeness employs Lipinski's rule of five, a widely used empirical filter to eliminate molecules that are unlikely to have oral bioavailability, as a key factor in determining whether a compound has a chance to become a useful drug. A compound passes Lipinsiki's rule rule and is considered druglike if the following conditions are met:
Molecules which fail this test are unlikely to cross cell membranes without the assistance of membrane transporters and are considered poor candidate to optimize into an effective drug. While most compounds that are eventually developed into drugs pass Lipinski's rule; however, some exceptions exist, particularly for compounds that are known to be substrates of membrane transporters.
Ki is the inhibition constant, the affinity of a ligand for a receptor. It is measured by obtaining the concentration of drug required to occupy 50% of the receptors and calculated with the Cheng-Prusoff equation using the IC50 value. Lower values correspond to more potent ligands.
Kd is the dissociation constant, describing how tightly the ligand is bound to the receptor. Lower values correspond to tighter binding between the two.
IC50 is the half maximal inhibitory concentration, which measures the concentration of drug required for 50% of maximal inhibition of function. Lower values correspond to more effective inhibitors.
EC50 is the half maximal effective concentration, which measures the concentration of drug required for 50% of maximal functional response. Lower values correspond to more effective agonists.
Potency is the measure of a functional response in terms of the concentration of drug required to produce an effect. Higher potency indicates that less drug is needed to produce a response.
Efficacy describes the way agonists vary in response while occupying the same number of receptors. An agonist with high efficacy exhibits maximal response while occupying relatively few receptors.
SMILES and InChI are unique molecular descriptor formats for describing the structure of a molecule using text. SMILES utilizes an ASCII string representation, as does InChI. A 27-character hashed version of an InChI ID is the InChI Key and was developed to expedite web searches. InChI has the advantage of being standardized, as SMILES strings for the same molecule can and have been known to differ; this is because SMILES is a proprietary format. However, both formats currently remain the most popular descriptors.
SDF is a chemical structure file format developed by MDL. Multiple structure entries can be stored within a single file, along with associated data, and are separated with a '$$$$'. Generally, the file provides atomic coordinates and connectivity of the ligand. PyMol is a popular molecular visualization system that can be used to visualize the ligands in 3D.
Chemical structure searching allows for the user to find ligands in the database by drawing and submitting a query molecule. There are two ways in which the user can search. The first is called substructure searching. Essentially, the query structure is used to see if any ligands in the database contain it as a part of the whole, hence the name. Using simplistic molecules for this search is discouraged, as many ligands may share the same substructure and cause the search to become very slow. Using a more refined query structure is recommended. The second is called similarity searching, in which the query molecule is compared with the entire database and are both represented as path-based fingerprints. Using the Tanimoto coefficient as a metric for similarity, target molecules are returned that match best based on a cutoff. Scores range from 0 - 1 and can be thought of as a percentage of similarity (i.e. 1 = 100%).
yangzhanglab@umich.edu | (734) 647-1549 | 100 Washtenaw Avenue | Ann Arbor, MI 48109-2218