July 28, 2014
"It was the first time in my career that I can remember, where creativity and outside the box thinking allowed for mixing and matching a lot of different and unrelated tools, to do something they weren’t originally intended, to generate a result that had immediate and positive effect on our research programs. Completely unpublishable work, I must admit, but...our own crystal structures would eventually come sometime later in the program, and when compared to our hand-made model of a model, we were very close, to the point where the crystal structure simply validated our model, which for a computational chemist, was intensely satisfying."
Dr. Cameron serves as Associate Director, Medicinal Chemistry, and nominal Head of Drug Discovery at viDA Therapeutics Inc., in Vancouver, BC, Canada. His professional experience over the past 18 years includes a variety of roles in Computational Chemistry, Medicinal Chemistry and Intellectual Property (IP) at small and large pharmaceutical companies and as a private consultant. From 2009 to 2012 he operated as a consultant to two companies in a variety of roles including computational chemistry, intellectual property, and business development support. Prior to that he served in several roles at local biotech company, MIGENIX (formerly Micrologix Biotech), most recently as Director, Research (Anti-Infectives) and Manager, Intellectual Property (in a concurrent role) and was responsible for Drug Discovery Research activities in antibacterial, anti-viral, and anti-fungal programs. He also managed the IP portfolio and other roles including evaluation of potential licensing opportunities. Prior to this Dr. Cameron spent almost 6 years at Boehringer Ingelheim (Canada) Ltd, in Laval, QC where he most recently served as the Head of Structural Research as a Sr. Research Scientist in Computational Chemistry. In this role he was awarded, as part of a team, the 2000 Boehringer Ingelheim R&D Award for his role in the identification of BILN2061, the first Hepatitis C Virus Protease inhibitor to proceed to human clinical trials. He is coauthor on over 30 scientific publications and is an inventor on at least 12 patent families. He has a Bachelor of Science (Hons., Subject of Specialization) in Chemistry and a Ph.D. in Chemistry both from Queen’s University.
Interviewed by Collaborative Drug Discovery, Inc.
As a community of researchers, how do you think we did with HIV versus HCV? And what, if any, are the lessons for future anti-infective drug discovery projects that we perhaps haven’t even yet come across (i.e. new pathogens that will emerge in the future)?
As a researcher who worked on both viruses, but much more heavily on HCV, I think we have done a great job. I recall the start of HCV research and how little we knew, but how much we relied, anecdotally, on the success of HIV to think about how to tackle HCV, even though there were quite a few differences between them. I’d like to think that the initial and continued success of HCV Protease agents and the future for other agents has provided support for a continued involvement in development programs for new pathogens. My hope is that the industry also heeds the many warnings out there and returns to the anti-bacterial arena as well.
Can you talk a bit about the viDA Therapeutics expertise, based on newly identified pathological roles for a family of granule secreted enzymes called Granzymes?
viDA Therapeutics Inc. is building on the expertise of Dr. David Granville from the University of British Columbia (UBC), primarily on Granzyme B (GzmB). His research has shown a role for extracellular GzmB in many areas, but most recently in photoaging, wound repair, and abdominal aortic aneurysm. Our goal is to affect tissue repair by specifically knocking out extracellular GzmB’s ability to interfere with the normal processes of healing. Our current focus is on a topical treatment for discoid lupus erythematosus (DLE).
How did your team at viDA Therapeutics come together?
The founder, David Granville, initially hooked up with our CEO Alistair Duncan, who had a background in the pharma industry as a CEO of Chromos Molecular Systems, and after numerous sessions at a local watering hole, viDA was born. The company has always operated somewhat virtually but since then has grown, as needed, to see an in-house biology lab, inside Dr. Graville’s lab at UBC and several others at our office location that keep the virtual side of the company pushing forward. The common goal is to develop an anti-GzmB drug and that goal keeps us all driven towards success.
On LinkedIn you mention being “a hands-on leader with experience managing both in-house and external CRO teams” – can you expand a little bit on what that means for you, as well as both why and how the CDD Vault helps, given your experience and work style preferences?
I’m a hand-on person, so no matter how far I get from the sciences I’m managing, I still like to get my hands dirty, literally when possible, but at least figuratively, so that I can get the best out of my team. I think it is important to build a team that respects each other, has a good understanding of the skill sets and limitations the team/facility has, and stays driven to the same common goal. This can be a challenge when portions of the company are in-house (and at remote locations), and portions are coming from contract research organizations (CROs) outside the company. For example, we contract out all of our synthetic chemistry to a Canadian CRO. Although all of the compound design occurs inside viDA, we treat our Chemists at the CRO as if they are part of the team, as best we can, keeping them abreast of the progress in the program and major advances going on at viDA, even if some of that is not required for their daily work. It’s important to me that all key members of our team (internal or external) have a good “feel” for where we are going, and the urgency to keep the research progressing.
CDD fits the bill for a virtual-based company very well. I’ve used other systems in the past that required not only a full-time IT support person to keep them running (especially after every new Microsoft Windows update), but also at least one and often more, science staff, in part-time roles, to support the science side of the software and to roll-out new templates, and methods, and such, as they needed to be implemented. On top of the infrastructure costs, this reduces productivity somewhat, as those folks are not innovating on the drug program, while they are performing simple support tasks. I didn’t want to waste a precious second of any of my team members, so with CDD, the IT expertise is centralized (and transparent) at CDD and I have almost instantaneous access to the scientific support structure for anything I can’t quickly solve on our end. Not to mention that we no longer need to spend some of our cash on infrastructure (due to the software-as-a-service model), our setup times for new protocols and loading times for new data are very small. This allows our team to share all of the data needed, in a seamless and easy environment (even remotely on smartphones), without being bogged down in difficult processes, or even to be tied down to their desk.
Because CDD Vault is web-based, I can easily use it whenever I am visiting one of our CRO groups off-site, or even virtually when I share my desktop, so they can see our data as needed, first hand, without the need for a lot of prep time. Although we don’t give our CROs access to the CDD Vault directly, it is still an invaluable tool for our hybrid operation here; I can’t imagine that we would have had the success we've had over the past few years, without the CDD Vault.
Why are some researchers better at computational work, and others better at experimental work? And how can you get the best of both worlds to accelerate real world drug discovery? In general, and especially in light of the new CDD Vision Module and “build models” capabilities within the CDD Vault.
That’s a good question. I see the two as being very connected. Both require conceptualization of the hypothesis you want to test, and both require intimate knowledge of the methods available to you, and the techniques needed to get the most out of the methods. Both also must have a good idea of the pitfalls and the flags that warn you if your analysis is valuable or misleading. So, in the end, I think it may simply come down to some things like “hands”. I know for me, experimental work always excited me, but I never found the right “feel” when doing the experiments. I was excited by the concepts and getting results, but performing experiments was not for me; my hands just didn’t lend themselves to that. I revel in the computational side, as I can put a bank of processors on a problem, work on other tasks and get my hands metaphorically dirty that way; all at the click of a keyboard button.
I think I’m in a somewhat ideal situation as a “jack of all trades”, as I have a good understanding of the experimental and computational side, and I see the value of mixing the two, as they both have their strengths, and the success of both is vastly improved when they are combined. Experimental work is sometimes only explainable when computational work supports one hypothesis over another. Certainly computational work is always going to be improved as well, when experimental knowledge can be integrated to focus in the area you need. Just like a good handyman, you’ve got to recognize the right tool for the job, use it effectively, and be willing to try new things to reach the desired goal.
I think the CDD Vault’s seamless integration of experiment and computation is a great start. Building a hypothesis that is based on experiment; and seeing how you can test the hypothesis computationally could be an eye-opener for those who live just in one of those two worlds. As we move forward, I see more and more mixed experimentalist/computational jack-of-all-trades emerging; especially as the tools to allow them to thrive become more mainstream and more companies embrace the new model.
Tell me about your research, current at viDA Therapeutics, and what you were doing previously?
My focus at viDA has been to develop novel, selective inhibitors of GzmB (a serine protease), that could be developed as topical drugs. To do that, we have a mix of computational chemistry, structural biology, biochemistry, medicinal chemistry, and synthetic chemistry, all working relatively seamlessly inside or outside the company, integrated with CDD, all motivated to meet our goals. Before viDA, I had the opportunity to be in a big pharma setting at Boehringer Ingelheim in Laval, a small pharma setting at Micrologix Biotech/MIGENIX in Vancouver, and more recently, a consulting setting, when I struck out on my own for a few years. The vast experience I got in those different settings, working on many different problems, has made me the researcher I am today. There is a good benefit from seeing things from all sides, and that has allowed me to really adopt the tagline “Do what you should do, not simply what you can do”. This may seem obvious, but it is intensely liberating when you have a feel for what this means to your daily work. At big pharma, as a bigger team with varied expertise, we all did what we could do, to build a very strong story. This, of course meant many resources being used without necessarily a clear path to success, and in some ways, conflict rather than teamwork, as individuals continued to do what they could do, even if it wasn’t affecting positive change in the program. At other small companies, you often are stuck with smaller teams with fragmented skill sets. It is important to recognize the minimum set of information required, and with a clear focus on that, you can best use the resources needed, to get what is needed done first (do what you should do). In a consulting setting, a little of “do what the customer wants” should always be mixed in, but it is important to always see what the best path forward is, and take it, so that your time and energy are not focused incorrectly, and you maximize your client’s investment to get where they need to go.
What is the end goal: modest and more audacious?
viDA’s end goals? I guess they’re like many other companies… make money for your investors, develop a drug that helps people, build notoriety and respect among your peers… the usual suspects. Our near term goals are more grounded. We hope that working on a topical drug development pathway, currently focused on DLE, that we can see proof of concept validation of our efforts in as short a time as possible, while spending the least amount of cash. With the new model of pharmaceutical research involving more small companies needing to do the early stage work, we think this is the best way to make progress and get noticed; which will be the best overall way to get GzmB inhibitors into settings where they can help people at risk to GzmB-related diseases. If we can make a difference in patient’s lives in DLE, it will be a launching point to others.
My own goals? They’ve been focused for some time now. Basically I want my professional life to have meant something, so if I can be a successful drug hunter (with an approved drug), that would obviously be the feather in the cap. However, the destination isn’t the whole picture. It’s the path you take to get there, and the people you share it with, that make for a rich life. So my immediate goal is to have fun, revel in the little successes, trudge through the setbacks and keep moving, while surrounding yourself with a team of people who think the same.
How do you use CDD Vault? Why?
CDD is the place I live every day. It’s the meeting room where we talk about and share new data. It’s the idea generation place when I’m looking for new structures. It’s really the foundation of our research program. Every assay or technique we embrace, is rolled into the CDD Vault at one time or another, so that we minimize the need to force people to see new data in meeting settings. Our team uses it when they need to, where they need to, for whatever purposes they need, so that we are as effective as we can be. Having all of that data at our fingertips is a real benefit and is helping to drive our programs.
What are the top 2-3 things you like about CDD Vault?
There are many but I think on top are things like:
- It is robust and flexible, while not being overly difficult to use – the handling of chemical structures is as robust as I could want, so that I know that new compounds with different structures are being handled correctly. At the same time, it is not so regimental that I can’t do certain things. So far I have found a way to do pretty much everything I want, and the cost of getting that to work is far lower than any other product I’ve worked with.
- Being web-based, it gives me complete freedom of when and where I can view the data, which allows me better access to my data, in a timely fashion, while not being tied to the traditional environment. As a member of a small virtual company that is a big benefit.
- The people at CDD are a huge asset. It is a big benefit to not only know, when you contact support, that your issue will be handled effectively and in a timely manner, but also that the support staff “feel your pain” and will do whatever it takes to get the answer to you, not matter how simple or difficult. To wrap all that together with a great personality and wonderful communications skills, makes for a pleasing all-around experience. I can’t say enough about working with Charlie and Sylvia on the support and business side of the business.
What are the top 2-3 things you wish were better in CDD Vault?
I can honestly say that the product has done nothing but improve over the time I’ve been using it. I see constant inclusion of ideas and requests that I’ve brought up, and the growth of the product has been spectacular. As with any simple to use SaaS product, there are bound to be limitations from time to time that are difficult to include, but for the most part I have very few things I’d like to see added at this time. Most of the things I see that could be better are very minor technical items, like being able to see all the calculated pKa values for a compound (not just the perceived strongest one), or to have two sets of dose response data for the same compound (on the same plate) not be combined to generate one IC50 curve, for times when you are testing things like effect of plate location on IC50… Having said this, these are easy to deal with in other ways.
If there was one thing I’d like to see in CDD Vault, it would be the ability to have two types of compounds in my vault; real and virtual (with two different coding schemes). I have always strongly felt that the main database of compounds that a company has, must be “real”. Meaning when I assign a certain code to it, I know that compound exists (or existed) at one time. Virtual libraries are becoming a huge tool for driving research but they can become vast very quickly, and in the absence of real data, they can simply become large libraries of little value. I will not allow virtual compounds to be given one of our corporate codes, so the use of Models is limited somewhat. I’d like to be able to use a separate code for virtual compounds so that I can a) test hypotheses inside and outside CDD Vault, while still maintaining a record of virtual compounds, b) make full use of Models and c) allow CDD Vault to be a collaborative tool for my external CRO synthesis target tracking. For example, when I design a compound, it goes into my old LICSS/Excel tracking system (see a question later on for more details on this) and we track synthesis there until it is made, then it goes into CDD Vault. If we had the capabilities of having virtual compounds, they could be tracked directly inside CDD, and make the whole process more seamless.
What was the most memorable interaction you've had w/ another brilliant scientist - and why?
I had the opportunity to meet Alfred Bader several times while at Queen’s University, doing both my undergraduate and PhD degrees. Alfred, the co-founder of Aldrich Chemicals (the standard for research chemicals for many years), was a Queen’s University Alumnus and great supporter, so he often came to visit us. Always enthusiastic about science, even late in his career, he also stressed the need to be well-rounded, and in that, embraced the arts almost as much as the sciences. He used to tell stories about the various cover art that the Aldrich catalog used to be covered with at the time, (and now emblazoned on the cover of Aldrichchimica Acta). His enthusiasm was not lost on me, and partly due to seeing him bask in the glory of creative things that weren’t “science”, I began to embrace my own creative side, and to this day maintain a healthy attitude about well rounded character, as I continue to be an amateur photographer, where I try to use my creativity in “other than science” ways, in the style of Alfred. He was the first scientist I recall knowing, who had as active a non-scientific life as his scientific life, and I always admired that.
What technologies (it doesn't need to be only about CDD) do you use for your research? Why?
Being part of a multidisciplinary drug discovery team means we have access to a large variety of tools that push our progress forward, although being part of a small, largely virtual company, we also have to live with the various limitations in our group’s capabilities which often makes it quite hard for us to take advantage of the “latest and greatest” techniques. Having said that, I don’t have time to address all of the tools we do have access to, but we do rely on key aspects as we push along. For example, computational chemistry tools, such as conformational modeling and virtual screening, and structural biology/research tools such as x-ray crystallography, are invaluable for new idea generation. It allows the chemistry efforts to be focused on ideas that look promising without needing, at every stage, experimental validation. Obviously we do prioritize different chemical series, and depend on experiment to drive success, so once a new compound is designed, we work closely with our synthetic chemistry CRO team to build the various compounds by a variety of methods. After synthesis, we depend heavily on assay technologies (fluorescence screening, 384-well plate handling, etc.) and in-house biochemistry, to feed data into the CDD Vault for the iterative process to start over again. There are so many tools used in that process that getting into the details would be the subject of a whole separate conversation. To be honest though, we are not on the forefront of new techniques or new technologies. Rather we use tried and tested tools to push our research project forward.
How do you work collaboratively within your group? Talk about your collaborations with others - what works well, what could be better?
The method of collaboration depends on the group and the functions needed. Being a virtual company for the most part means we have many pieces that are not inside viDA and that means fruitful collaboration requires seamless communication, using whatever tools we can. The CDD Vault is the base of some of those communications but many other tools are used as well.
Ideally, the best tools are the ones that are easiest to use and contain the data needed to collaborate, while having the most control over data security, while maintaining an environment that is simple to administer and affordable.
Talk about an 'ah-ha' moment in science or business during your career?
I’ve been privileged to be a part of many 'Ah-ha' moments in my career because I’ve spent a long time in very active research groups. Many of these revolve around finding the link between two things that just wasn’t clear until some triggering event. I’ll tell you about one that made a difference to me early on in my career.
Early in my career at Boehringer Ingelheim, at the beginnings of our HCV Protease project (mid- to late-90’s), we did not have crystal structure information for the target of interest. Agouron (now part of Pfizer) had published the structure, but at that time the structure was not required to be released for public use upon publication, so the actual 3D coordinates were not yet publicaly available (and would not be for a year). Given the large variation in sequence between the protease and other proteases with similar folds, we felt just doing homology modeling with that many sequence variations would not help. It was then that I realized that the alpha carbon trace of the Agouron structure had been provided (in a stereoscopic view) in the paper. Since the distance between alpha carbon atoms is relatively fixed, using a digitizing method, I could take the alpha carbon trace, and rebuild it in our modeling software (generating 3D coordinates from the 2D picture) and using the other various images provided, we could effectively place the key sidechains and amide bonds in relatively correct positions, so that we could devise a model of the key active site to begin docking our own compounds for hypothesis generation. We used to call it the “model of a model”.
It may seem a little strange for something so simple to be an 'ah-ha' moment, but it was the first time in my career that I can remember, where creativity and outside the box thinking allowed for mixing and matching a lot of different and unrelated tools, to do something they weren’t originally intended, to generate a result that had immediate and positive effect on our research programs. Completely unpublishable work, I must admit, BUT it was the only tool we had to combat our competitors that were “ahead of us” at the time. The end result of this was our team’s ability to push BILN 2061 to be the first HCV Protease inhibitor tested in humans, where we showed that you could drop the viral load, dramatically and immediately, in humans. We published this work in Nature, and although BILN 2061 did not make it to commercial release, it was a validation of the concept, and was an inspiration for others in the field to keep innovating; and in fact, key structural discoveries observed in BILN 2061 exist today in the protease inhibitors that did make it to commercial release. Our own crystal structures would eventually come sometime later in the program, and when compared to our hand-made model of a model, we were very close, to the point where the crystal structure simply validated our model, which for a computational chemist, was intensely satisfying. That creativity remains with me today and pervades into my research regularly.
Outside of your own research (it can be a competitor or an unrelated group), what is the most fascinating development or study you've recently seen, something that would be broadly of interest to the researcher community?
I think the most interesting things I’ve seen lately all come from last year’s Gordon Research Conference on Computational Chemistry which was run by Anthony Nicholls of OpenEye Software. The goal of that conference was a little different than usual, and he focused the entire program on things like reproducibility, error handling and statistics/validation, so that the experimentalists and computationalists could have a better understanding of each other’s “worlds”, and build better bridges between the two. I’ll share one effort I was not aware of until that conference, which I think is going a long way towards making our industry, and the academic world that supports it, overall more accountable and useable.
Elizabeth Iorns, CEO of Science Exchange, spoke about something called the Reproducability Initiative, which was focused on offering ways to have key studies be reproduced in independent labs to validate key observations. This is critical in our new day and age, where the focus is to keep moving forward, not to take a look at what you’ve done and be sure you haven’t made errors. The reality is most academic researchers are rewarded for publishing, not validation of published work. And as such, there is fast becoming a disturbing correlation with journal impact and number of retractions, and this has led some to say that at least 50% of all academic research today is unreproducible. See this link for some of the common rules for studies that are not taken into account when designed, carried out or published.
I think this Science Exchange is a wonderful advance in science, as it is a way to offer validation services to almost anyone, in a seamless environment so that you can be sure that observations in the literature are real and can be trusted, before embarking down what could be a wild goose chases, or worse yet, not embarking on what could be excellent prospects. This kind of initiative is a great demonstration of how scientists want their fields to be taken seriously, and by addressing validation, on a case-by-case basis, they are making the whole community more accountable and accurate. Although not a huge scientific discovery or new technique, I think this initiative, and others like it, will revolutionize the way we do science moving forward.
Before you used CDD Vault, how did you manage your data or collaborate?
Before CDD Vault, I used a mix of methods to store data. As we are early on, Excel proved to be the most useful for the biology data but we did also use Prism to calculate summary data like IC50 values so it meant a lot of transcription, which can be error-laden and limited in terms of lack of real-time update. These summary tables were shared using SharePoint, or Microsoft 365, so that everyone on the team could keep up to date with progress; these sharing tools had ways to inform staff when the base file had been updated so they could have notifications of new data “pushed” to them as needed.
For Chemistry I always had a molfile for every structure, drawn in a standard way, so we could load data into a future solution. For day to day tracking, we used Excel with a plugin called LICSS, which was a simple way to add some chemistry functionality to Excel at no expense, using SMILES strings to store the structural data in a “future-ready” format. I shared this “chemistry-aware” tracking Excel sheet with our Chemistry CRO team (using shared internet storage options like DropBox) so that we not only knew what we had made, but what we planned to make as well.
This combination of methods allowed us to have a good level of collaborative abilities, while still maintaining a secure hold on our data. The downside was the large amount of labour it took to maintain these various tools. CDD Vault really took all of that collaborative ability, rolled it into a better package, added more functionality and made it simpler to use and maintain.
Why did you choose CDD Vault? In retrospect, why was it a good or bad decision?
Having used many tools in the past, I wanted something that had the rigor I needed, while having a real cost (licensing, equipment, staff, etc.) that was not prohibitive, as we were and continue to be a small virtual company with a need to minimize infrastructure. CDD Vault, with its SaaS model, was by far the best possible solution for me and after meeting with the team and seeing the rigor that was possible in the product while seeing the flexibility…I was sold. CDD Vault was definitely a good decision, if you ask me. It’s part of the reason we’ve been as successful as we have been and while being able to stay focused on our goals.
This blog is authored by members of the CDD Vault community. CDD Vault is a hosted drug discovery informatics platform that securely manages both private and external biological and chemical data. It provides core functionality including chemical registration, structure activity relationship, chemical inventory, and electronic lab notebook capabilities!
CDD Vault: Drug Discovery Informatics your whole project team will embrace!