You are here: Services >> Services

Services

Details of the services that Equinox Pharma provides

Services

Overview of services

Equinox Pharma offers a customised, unique, and novel hit and lead molecule finding service for pharmaceutical, biotechnological and agrochemical customers on a fee-for-service basis using its patented and proprietary INDDEx™ software and a set(s) of biological screening data (data sources). These input screening data can be proprietary to our customers or originate from the public domain.

Equinox Pharma will provide, to its customers, an intelligible interpretation of complex screening data, and chemical rule-based models which relate structure to activity. These models are expressed in forms readily understood by medicinal chemists. The intellectual property behind these models will be owned by the customer.

These models will be used to screen either public-domain or proprietary databases of compounds to predict a ranked list of active compounds. Importantly, the INDDEx™-derived models identify active compounds of different structural types from the input screening data (scaffold-hopping).

Equinox Pharma can, on a service basis, take these hits and perform hit-to-lead optimisation using further iterations of INDDEx™, commercial acquisition of compounds and outsourced lead optimisation services. Equinox Pharma can also intelligently interpret complex screening data, for customers, providing readily comprehensible rules relating substructure to activity, and identifying different binding modes.

In addition to identifying and optimising potency properties, it is important to note that the INDDEx™ software can be employed to identify and minimise other properties, for example, toxicity and ADME liabilities.

Equinox has worked with several such customers in successfully employing these services and as a result, identifying novel series of potent, ADMET-friendly, optimised leads for further development.

Input screening data

For the discovery of novel molecular hits and leads for its customers, Equinox Pharma requires the ligand activity and structure of a set of data from a screening programme as learning input for our proprietary INDDEx™ software. The number of active compounds in the dataset is preferably in the tens to hundreds, but we have demonstrated success with considerably fewer or greater numbers.

Biological activity can be provided either in a quantitative (e.g. logKi) or qualitative (active/inactive) form. For the best hit and lead discovery results, it is recommended that a quantitative measure be used. This is provided in the form of a measure of activity: usually a measure of ligand affinity (logKi, nM, μM), but other measures of potency (% enzyme inhibition, log[kg/hectare], etc.) can also be used. Hit and lead discovery can also be performed using a purely qualitative method, in the form of a list of actives and inactives without any measure of potency.

INDDEx™ derives its logic-based rules from 3D molecular structures. When customers provide these structures (which will be kept under the strictest confidentiality), they can be either in the form of SMILES strings, from which Equinox Pharma can create energetically minimised structures, or in the form of 2D/3D common chemical file formats such as SDF or MOL2. List of formats currently supported.

Rule-based model

INDDEx™ learns positive rules from active molecules in the dataset, but also learns, and indeed requires, negative rules from inactive molecules. A list of molecules known to be inactive can be provided by the customer, or Equinox Pharma can select molecules from an in-house dataset of molecules for INDDEx™ to learn negative rules. INDDEx™ can learn particularly powerful rules when the screening dataset is structurally diverse and yet when the inactive compounds share structural similarities with the active compounds.

The outcome of INDDEx™ is a chemical rule-based model which relates structure to activity. This model is paid for, and can be exclusively owned, by the customer.

Predict a ranked list of active compounds

The rule-based model(s) is used to screen in silico databases of chemical structures to predict new hit structures. These databases of chemical structures can be publicly available or proprietary to the customer or other third parties. The structures can be known or theoretical. Equinox Pharma typically employs the ZINC database which contains the 3D minimised structures of over 14 million compounds that are marked as being readily purchasable.

Equinox Pharma will produce a ranked list of predicted active compounds from the database employed in in silico screening. Equinox Pharma can then purchase those identified compounds and screen them for actual activity. This screening can be performed by the customer or Equinox Pharma can outsource this activity. A list of actual activity versus predicted activity can then be provided to the customer.

For best results we would envisage cycles of prediction, testing, and refinemenent, as molecules with high predicted activities are tested, and the results used to improve the model developed by INDDEx™ and the quality of the hit and lead series provided to the customer.

Equinox Pharma can perform discovery of new hits and provide results, or it can outsource primary and secondary screening to a third party.

To discuss your requirements in interpreting screening data, lead discovery and lead optimisation, using INDDEx™, please contact us.

 

Data sources

Sources: In-house / public domain / purchased from a third party (or combination).

Possible sources of ligand activity data
In-house The customer’s data from bioassays conducted in-house. Equinox Pharma does not require any knowledge of the enzyme/receptor target.
Public domain Equinox has experience using public domain bioassay data, such as data provided by PubChem. Advantages are that the data is totally free and immediately available. The disadvantages are that there is a limited choice of assays, and the data may not be as clean as in-house or third party data.
Purchased from a third party Data can be purchased from third-party companies who conduct/have conducted bioassays.
Combination A combination of any of the above three.

 

 

Chemical file formats supported

alcAlchemy file prepAmber PREP file
bsBall & Stick file caccrtCacao Cartesian file
cccCCC file c3d1Chem3D Cartesian 1 file
c3d2Chem3D Cartesian 2 file cmlChemical Markup Language file
crk2dCRK2D: Chemical Resource Kit 2D file crk3dCRK3D: Chemical Resource Kit 3D file
boxDock 3.5 Box file dmolDMol3 Coordinates file
featFeature file gamGAMESS Output file
gamoutGAMESS Output file gprGhemical Project file
mm1gpGhemical MM file qm1gpGhemical QM file
hinHyperChem HIN file joutJaguar Output file
binOpenEye Binary file mmdMacroModel file
mmodMacroModel file outMacroModel file
datMacroModel file carMSI Biosym/Insight II CAR file
sdfMDL Isis SDF file sdMDL Isis SDF file
mdlMDL Molfile file molMDL Molfile file
mopcrtMOPAC Cartesian file mopoutMOPAC Output file
mmadsMMADS file mpqcMPQC file
bgfMSI BGF file nwoNWChem Output file
pdbPDB file entPDB file
pqsPQS file qcoutQ-Chem Output file
resShelX file insShelX file
smiSMILES file mol2Sybyl Mol2 file
unixyzUniChem XYZ file vmolViewMol file
xyzXYZ file