Consortium Offers Compound Library and Screening Facilities to UK Researchers

UK Announcement

A very unusual, and beneficial, research resource is now available to UK researchers:  a world-class compound library and associated high-throughput screening facilities and expertise. The resource is larger and more targeted than most comparable resources, such as the NIH compound repository. This resource can be used for any biomolecular screen; such approaches are not restricted to drug discovery. This resource may be one that can stimulate research collaborations. UK has taken considerable care with respect to IP terms; use of the library is very unencumbered by IP restrictions. The bulk of the compounds are known, commercially-available molecules that are very well-characterized.

UK has joined with the University of Cincinnati Genome Research Institute and the Research Institute of the Children’s Hospital of Cincinnati as members of a consortium to make available to our researchers and research collaborators a drug-discovery compound library, high-throughput screening technology, and expertise to use the library, which we call the UC/GRI Compound Library. The library was recently donated to the University of Cincinnati for consortium use. This library was carefully designed and maintained in order to provide the ability to screen biological targets for novel biochemical tools, probes and drug leads. This memo links to basic information on the library, links to presentations by UC/GRI that describe the library and their facilities and expertise, and the guidelines for UK researchers wishing to use the library.

Guidelines for UK Researchers

The Office of the Vice President for Research funded consortium access. Initially, to encourage screening, we will provide up to 25,000 compounds to a UK researcher who develops a screening plan with UC/GRI, essentially as a cost-free minigrant.  Larger or additional screens may be done at a subsidized cost. The actual cost of screening assays, etc, will be borne by the researcher. A UK researcher should read the associated library description, contact UC/GRI’s lead on this project, Dr. Bill Seibel at william.seibel@uc.edu, develop a plan, then submit a very brief description of the plan to Associate Vice President for Research Chuck Staben for approval. Response to requests will be vetted in our office with a very quick response time.

Links to UC/GRI PowerPoint Presentations

• Introduction to HTS Screening-general information (ppt)
• Introduction to UC/GRI including information on HTS facilities, introduction to library (ppt)
• UC/GRI Chemical Library-more detailed library information (ppt)

Compound Library Overview
The main UC/GRI Compound Library is comprised of ca. 250,000 compounds designed to be an optimal collection for screening of biological targets.  The overriding concepts were simple:

1. The compounds should be restricted to those reasonably conforming to drug-like properties.

   Drug-like space was determined by characterizing the World Drug Index and only purchasing compounds from vendors that fit closely around and within this space. 

2. The compounds should span this drug-like space as much as possible to ensure the potential for hits against any target that is screened.

   Computational methods were employed to ensure a relatively uniform distribution of compounds across this space, avoiding an over-concentration in any one zone.

3. Compounds containing structural features that lead to poor quality hits have been excluded from consideration for purchase.

   A number of selection criteria were developed by screeners and medicinal chemists to act as filters. Approximately 80 reactive or tox-associated functional groups have been identified in addition to broader descriptors such as logP, predicted logD, Lipinski, and Veber rules to ensure the collection avoids compounds of low potential for advancement.

4. It is recognized that some drug targets are of greater medical relevance than others, and a slight emphasis is given to those areas. 

   Small libraries directed toward kinases and GPCRs were added to the collection. In addition, the company which assembled this collection included many of the compounds prepared as part of its research programs and included kinase, GPCR, protease, metalloprotease, phosphatase, nuclear hormone receptors, and other common medically relevant target classes.

This library was designed by a group of screeners, medicinal chemists, computational chemists, and assay specialists with extensive collective experience in the discovery of drugs in a mid-sized pharmaceutical company. The result is a carefully constructed library of compounds that is broad across drug-like space, with concentrations in medicinally relevant areas. 

Beyond the screening collection are a variety of other compounds that may or may not conform to library inclusion criteria, so while our primary screening collection consists of slightly over 250,000 compounds, the entire collection contains over 340,000 compounds, which can be accessed if appropriate. Structure-Data (SD) files of this collection are available to consortium members for their use. Consult with Dr. Bill Seibel at william.seibel@uc.edu for access to this structure data file. We request that this file not be distributed outside the consortium without a confidentiality agreement. 

Physical Library
The Compound Library is currently housed in 3000 square feet of laboratory space at a separate research facility in Mason, Ohio. Neat compounds are stored in a Haystack™ automated storage and retrieval system, which has a capacity of ca. 200,000 samples. The majority of compounds were purchased as 10 mg samples. Compounds are dissolved to 10mM concentration in DMSO and stored under Argon at 4°C and 0% humidity in a Solar™ automated storage and retrieval system, which has a capacity of 1,800,000 samples in microtiter plates and microtubes.     

Screening with the UC/GRI Library
Compounds exist in rather limited supply, so it is imperative that consortium members optimize assays and library selection for the most efficient use of our compounds in addressing the key project needs. Potential screeners should discuss their screen and the optimal use of the library in their screen with Dr. Sandra Nelson and Dr. Bill Seibel at UC/GRI. Dr. Nelson has 20 years of experience in the development of assays and use in screening programs. Dr. Seibel has 19 years of experience in medicinal chemistry, library design and hit follow-up/lead optimization. The University of Cincinnati’s Genome Research Institute has a state-of-the art High-Throughput Screening center, featuring an Evotec Plate Explorer system with an Opera™ high-content reader, capable of most high throughput and high content assay readouts. Consortium members are entitled to preferred screening rates with this system. 

For any screening program, three main approaches to library construction are feasible:

1.  Diverse collections

• Comprehensive screening against all available compounds (ca. 250,000 cmpds)
• Screening against a representative subset of available compounds (5K, 10K, 15K, 25K cmpds)

2.  Class-associated compounds

• Compounds with structural features often associated with the target class of interest (e.g. kinase-directed)

3.  Structure-based compound selection

• Compounds identified from a Virtual Screen of the target crystal structure (or high quality homology model) to select most likely inhibitors (e.g. 500 cmpds)
• Develop pharmacophore models from known ligands of the target and select compounds from the library based on fit to this model. (e.g. 500 cmpds)
• Conduct broader “similarity” searches of the library based on one or more ligands (e.g. 1000 cmpds)

We are certainly open to other approaches, and happy to discuss any options. 

Compound Access
The compound library was constructed primarily for screening purposes and simple confirmatory/analytical work. As such, compound samples are generally 10 mg or less, and we all need to do what we can to conserve this resource.

Consortium members have each contributed funds toward maintaining the repository, and this amount roughly covers the cost of distributing an allocation of 250,000 samples, which each consortium member organization is entitled to at no further cost. Please contact Chuck Staben after you have devised your access/screening plan concerning how to access the library. Additional samples are available at cost to consortium members.

Compounds will be supplied as 5 µL of 10 mM solutions in DMSO in either 384-well plates (other options can be discussed). Requests for more compound than this are discouraged in order to preserve the compound supply. In general, users should anticipate purchasing follow-up samples from a commercial vendor. We are happy to provide this information to collaborators. For compounds that are not commercially available, we can help find synthetic protocols and/or identify organizations who can help with the synthesis. 

Compound Requests
Diverse subsets have been designed and are conveniently available in 5K, 10K, 15K, and 25K subsets. If researchers wish to conduct their own virtual screening or similar protocols, SD files of the library have been made available to consortium representatives for distribution within their organization as appropriate. We can provide this file as well on CD. 

• Requests for compound sets should be made via eMail to Dr. Bill Seibel at william.seibel@uc.edu. 
• For specific compounds, send the compound numbers in an Excel or text file. 
• For preformatted library requests, simply state the preferred library desired in the email (e.g. 15K diverse subset). 
• Compounds will be supplied in sealed 384 well plates at 5 µL of 10mM/well, shipped in dry ice containers. 
• Plate maps with ID number and Batch number will be supplied in Excel spreadsheets. 
• Requests for confirmatory samples of screening hits should include the compound number and batch number, to ensure the same material is used.
• We expect to be able to ship most requests within two weeks of each order receipt. 

Hit Follow-Up Activities
Hit Confirmation:  We highly recommend close scrutiny of the initial screening data, particularly confirmation of activity and confirmation of sample identity and purity (e.g. LCMS). Although our collection is stored under the best possible conditions, compound decomposition is always a possibility.

Hit Expansion:  Implicit in the practice of screening a small subset of the repository is the expectation that one would search for compounds related to the hits identified.  This can be done via three methods:  Similarity search (e.g. Nearest Neighbors, Shape Matching); Substructure searches, and Pharmacophore searches (assuming sufficient info from the hits). These can be considered cycles in an iterative process as depicted below. Ultimately, the process can be extended beyond our in-house library to search the millions of compounds commercially available and purchase compounds for testing. This last cycle is also appropriate for the hits emerging from a full library screen. We can conduct these computations in collaboration with researchers if so desired.

SAR Analysis:  We have several medicinal chemists on staff who can assist in analysis of data sets for preliminary Structure-Activity Relationships, and may be available for extended consultation with follow-up programs. 

Medicinal Chemistry:  Our Med Chem staff is currently limited, but is available for consultation regarding follow-up activities, logical next steps and potentially helping set up collaborative opportunities. We may be able to undertake limited lab activities to help push the research along. We will document and notebook all information appropriate to assisting in legal filings supporting patent applications and or publications.