Modern drug discovery exploits high content, high throughput, and high complexity processes.
Managing the wealth of chemical structures, bioactivity and other assay data results requires an informatics infrastructure that puts data at scientists’ fingertips, and facilitates searching, analysis and data partitioning, comments Barry Bunin, president and CEO at Collaborative Drug Discovery (CDD).
‘The drug discovery space needs intuitive software that keeps data secure and accessible for a distributed set of authorised personnel. As collaborations evolve, new users need to be productive quickly, easily accessing shared research and results. Capturing data and experimental workflows in a single system provides the necessary context to analyse results and helps to avoid unnecessary repetition of experiments.’
Based in Burlingame, California, and Cambridge, UK, CDD was spun out of Eli Lilly in 2004 to develop and commercialise a holistic informatics platform that addresses the needs of the drug discovery sector.
The firm’s flagship CDD Vault is a hosted, web-based data repository through which drug discovery teams can access, mine and interrogate their chemical structures, assay results and associated experimental and research results. Built from the ground up as a single platform rather than an aggregation of software modules that sit atop a traditional, sample-centric laboratory information management system (LIMS), CDD Vault provides a full range of capabilities from a single login and integrated web interface. Functionality is organised into meaningful feature sets with no hidden add-ons: Activity & Registration, Inventory, Visualization and ELN.
‘Our founders really understood the cheminformatics space, and wanted to address the hugely underserved academic and biotechnology drug discovery sector,’ Kellan Gregory, CDD’s director of product excellence, explains. ‘Many of our competitors went after big pharma, developing complex, custom solutions that take years to set up and are expensive to maintain. Instead, we created a far more cost-effective, focused, and easy-to-implement platform that requires no IT expertise on the part of our customers.’
CDD Vault provides a dynamic visualisation tool to easily spot trends, zoom and filter to further refine, drill down on points of interests, and export to share with colleagues. The importance of visual inspection can’t be overstressed, Gregory suggests. ‘We all like pictures and it’s a necessary step of data analysis. We have built an amazing visualisation tool that leverages modern reactive design principles. Users can retrieve data from anywhere within CDD Vault, creating a report which is the foundation of a visualisation session. Compare up to 16 different parameters simultaneously, look for correlations or identify compounds of interest. Easily generate charts and graphs for presentations and reports.’
CDD Vault also offers key electronic laboratory notebook (ELN) functionality, developed for the drug discovery industry. The ELN captures research without the need for paper logbooks, keeping data searchable and accessible, and supporting collaboration with external and in house teams. Research can be documented using pre-defined fields according to business rules or as free flowing body of text, tables, files and images. Importantly, the ELN integrates seamlessly with the chemical and biological utility of CDD Vault, Gregory notes.
‘ELN is a bit of an overloaded term these days, and when you ask what people expect from an ELN they may want something that is either more like a LIMS or that facilitates unstructured data management, a true virtual notebook page. CDD Vault offers ELN capabilities that are tailored to the needs of drug discovery researchers, and we are actively developing new features that are immediately available to all our customers. We’re in constant contact with our users, gathering feedback after each release to further refine the usability and utility.’
CDD Vault’s biological post-processing and data analytics includes normalisation, concentration or dose response curve fitting, and an Excel-like formula builder for calculations such as BEI or ligand efficiency. ‘We realise that there may be occasions when labs need specialty tools, such as more advanced analytics or specialised visualisation,’ Bunin continues. ‘We have made it easy to pull data out of CDD Vault into external software, and then deposit the findings back into CDD Vault. As a central repository, the platform can capture and handle any type of data, even a file generated externally, a capture it in line with the originating data.’
The CDD platform is project-structured to facilitate data sharing and collaboration, and today CDD Vault represents a central data repository for some of the world’s largest collaborative R&D networks. ‘As far back as 10 years ago when the firm was focused on supporting drug discovery for neglected diseases, public private partnerships represented a major research driver,’ Bunin states. That PPP model has grown, and CDD Vault supports key global initiatives, including the Bill and Melinda Gates Foundation’s TB Drug Accelerator program, which includes eight pharma companies and seven research organisations, and the National Institutes of Health’s Blueprint for Neuroscience Research. ‘In this case the NIH is the CDD Vault ‘owner’, and contributors have access to individual projects. NIH can then coordinate, view and analyse data across all projects within the network.’
CDD Vault has been developed as a hosted solution from its earliest incarnation. ‘That was a major differentiator for us back in the early days,’ Bunin points out. ‘We have built a REST-based API, which means that just about anyone can connect to our system. Most recently we connected with Data Warrior via the API. You can launch a new Data Warrior session, select CDD Vault as the data source, and then pull data directly from CDD Vault into Data Warrior.’
Critically, CDD Vault is inherently easy to implement, unlike the ‘super-heavy custom solutions’, which are expensive, difficult to maintain, and require IT staff and engagement with a service component of the vendor. ‘In contrast, we manage all of the infrastructure, and can get a new customer up and running in minutes. Users can immediately import structures and assay data, and start making full use of the platform’s features. Our early users, including the University of California, San Francisco (UCSF) and the Broad Institute, were instrumental in the development of core functionality, including the data import wizard, and chemical and biological registration.’
CDD provides all customers with comprehensive support, training and troubleshooting as well as helpful online training tools. The firm encourages its customers to ask for help in whatever way they feel comfortable, whether it’s a scheduled phone call, an email, or a web meeting. CDD is also constantly refining and updating the platform. New releases and fixes are put out early and often, without customers having to endure any system down time. The philosophy is to refine the product as a result of customer feedback, which drives the development of upgrades and new functionality. ‘CDD differentiates itself from other providers by creating a product built for the entire ‘community’ perspective, instead of just a single large customer. Consequently, we retain high engagement with customers, including a remarkable >90 per cent renewal rate.’
CDD asks its customers to demonstrate how they are using new functionality for each newly released feature set, to see it working from different perspectives, and assess usability first hand. ‘We can then carefully evaluate if and when people are stumbling in the user interface, for example,’ Gregory explains. ‘Perhaps the steps required aren’t intuitive enough. By seeing how our customers navigate their daily tasks using CDD Vault we can continue to iterate and create a platform that is currently used on every continent, except Antarctica.’
CDD is in parallel working with customers in key areas such as biologics. ‘We really need to engage the community to help direct development of new functionality for the biologics discovery field,’ Bunin notes. ‘There are already a ton of tools out there, but there are also many gaps. We need to understand how we can bridge those gaps. We have assembled a fantastic team of scientists who have decades of experience, and we have no shortage of ideas. But it’s only by working with our customers that we will be able to prototype those ideas and test them as a core part of CDD Vault, along with complementary new capabilities.’
