Targeting Your Best Clinical Candidate With Self Guided Data Discovery and Analysis

We spoke with Daniel Weaver, Senior Product Manager for PerkinElmer, to find out more about their new platform PerkinElmer Signals™ Lead Discovery (SLD), powered by TIBCO Spotfire analytics. Daniel recently presented a webinar on Signals Lead Discovery, which you can watch here.

Ruairi Mackenzie (RM): What do you think are the biggest advantages of using a data platform for biomedical researchers working in candidate discovery today?

Dan Weaver (DW): Most scientists are still dealing with the same issues that they have been dealing with over the last 15 years I’ve been in the industry. The problem is that they spend a tremendous amount of time just assembling the data to answer questions. Previous solutions are starting to show signs of age, being built on two-tier client servers behind-a-firewall architectures That’s just not the way computing is done today and is not what the next generation of scientists expect. Despite that, essentially all the solutions that are currently in production, are still two-tier on-premise solutions and there is a gradual industrial trend to an architecture that is cloud enabled, readily available on multiple devices and is basically not old architecture. We are building towards that as the future, and PerkinElmer Signals™ Lead Discovery is a significant step in that direction.

The time is right for the next generation of solution.

SLD is attempting to move in the industry’s direction – giving scientists a much better way to find data in those new directions.  The advantages are that scientists can find the data that they need to find to answer scientific questions, as opposed to spending all their time assembling, reassembling and managing excel spreadsheets. 

RM: What are the headline features of SLD? What would new users of SLD want to know about it?

DW: The two most valuable aspects of the SLD platform are – self guided data discovery and self-guided data analysis. Our goal has been to make it easy for scientists to search over the entire catalogue of data that might be available to them at the corporate level and from that catalogue select the data that is of interest to them – the concept of self-guided data discovery. SLD aims to create an infrastructure where scientists can check off assays of interest and can easily express a set of interests in compounds and then get their data into a simple tool for conducting data review analysis.

The second part is self-guided data discovery. PerkinElmer is the exclusive life sciences R&D franchise owner for TIBCO Spotfire and we have had this product in the field for many years and we have gotten feedback that people love the tool. With SLD we have taken the capability set for TIBCO Spotfire and created a set of interfaces that are simple for a new user to be introduced to and get started on, whilst enabling self-confessed ‘Spotfire Junkies’ to conduct deeper analysis if they want. It’s still Spotfire, and is still full-featured, but the experience of it is simple and straightforward to enable scientists to begin answering questions as quickly as possible.  

RM: In the webinar you explored some of the modules and integrations that can be added to SLD. Could you talk us through some of these modules?

DW: In SLD’s platform, we have included some additional capabilities that go over and above the simplified headlines of Self-Guided Data Discovery and Data Analysis. I’ll talk you through three. One is the implementation of high-performance chemistry-based searching. This capability enables a you to search over an arbitrarily large set of compounds with basically predictable performance. So, we have taken the concept of chemistry searching and have parallelized it and put it into a big data back end such that if you have a million compounds or one hundred million compounds we don’t care. We can give you back a guaranteed search performance. Of course, with larger compounds you have to spool up more nodes but that’s the whole point of big data systems. 

A second important capability is a whole new set of tools enabling structure activity relationship (SAR) analysis. Those toolsets include the ability to visualize R-Group decomposition results efficiently and the ability to visualize 3-D structures inside Spotfire. For example, you can take a set of small molecules and do docking experiments and then bring those results and display them directly inside Spotfire. 

The third capability is that for the first time ever, we can analyze biological molecules. In addition to tradition small molecules we can now look at the capability or characteristics of large molecules such as monoclonal antibodies or CRISPR guide RNAs. We’ve really tried to broaden the applications of this tool across different scientific disciplines.

RM: With that in mind, is SLD a platform that can integrate various 'omics? 

DW: Not quite. This is something which is commonly brought up around SLD. Let's talk about what SLD is and what it is isn’t. SLD is about lead discovery, which in pharmaceuticals research is the process by which you take a set of biomolecules and try to identify which compounds within those molecules are your leads to advance to more expensive or complex experiments. So SLD's purpose is to enable a project team to more quickly and correctly identify which drug candidates to advance. For general data integration and analysis, PerkinElmer uses Signals Translational. These are two different products, with different aims and it is important to be clear that there is no one big mix. Just because we can handle biomolecules doesn’t mean we are integrating multiple 'omics data.

A model use of SLD for, say, a biologics engineer or protein chemist might be to assess a monoclonal antibody’s efficacy against treating hybridomas. In that process the scientist might want to look at things in for example the reasons for those variabilities - which amino acids correspond to a better result - SLD is all about looking at either R-Groups, small molecules or amino acid constituents on large molecules. Or, in this case, which modified amino acids are correlated with the best therapeutic efficacy. 

RM: What capacity does SLD have for cloud integration?

DW: The architecture that SLD is built on is what I’d call cloud enabled. The industry is still in a transitional phase, where they’re exploring using cloud-based technologies, but are hesitant to put data outside of their firewall. SLD is therefore installable in a cloud environment, for which we utilize Amazon Web Services, but also support clients who want to install inside either their own firewall or private virtual cloud. We’ve leveraged our technology partners Attivio’s experience in taking data into cloud environments. At the most fundamental level, the technology here is based on a leucine index, a big data, cloud-based technology. This really enables PerkinElmer to become more proficient with merging big data and cloud base technology that has come out in the last 5-10 years. We have also implemented interfaces from previous form-based tools, our goal being here to provide a path by which people are used to looking at data in those kinds of form-based views would have a comfortable enough transition into these new computing environments, which are fundamentally not form-based computing environments. All this together means we can produce a future-proof product that stills caters for scientists more accustomed to older technologies. 

Want to learn more about PerkinElmer Signals™ Lead Discovery? Watch Daniel Weaver's webinar here.

Daniel Weaver was speaking to Ruairi J Mackenzie, Science Writer for Technology Networks