Turning Structural Data Into Knowledge:GSK at the frontline

Gain easier access to data and you’re on the way to a smoother informatics Lab environment. GSK is on this journey and has recently implemented a piece of software that makes it easier for the user to drill down into the data and search from almost any facet. Fresh from the 7th Annual Smartlab Summit, Steve Thomas, Investigator at GlaxoSmithKline, speaks to  Pharma IQ.

Pharma IQ
: Steve, can you begin by giving us an overview on your role at GSK and the work that you’re doing in laboratory informatics.
S Thomas: I’m an analytical chemist. I’m described rather kindly as being analytically bilingual; I speak both mass spec and NMR. I work in a department called structural ID. It’s our job for GSK to look at the fate of all our molecules in the body so that we understand what the body turns them into to ensure that we don’t have a problem in the clinic with a metabolic liability. We have a lot of information coming through the department that we have to get right and we have to get it right quickly so that we can put the best and safest drugs out onto the market and improve peoples’ quality of life.
Pharma IQ: And what work are you doing with your laboratory informatics?

S Thomas: An awful lot of the metabolic changes that the body is able to do are similar for different molecules. And so, there is an element of not wishing to reinvent the wheel. What we would like to do is to be able to store, search and share our knowledge of the metabolites and the outcome; not just the data. We can use this knowledge to add to our corporate wisdom. If I have a metabolic result that would help a colleague of mine in the US understand his data, I will want them to be able to have access to it. Equally, if I’ve got a colleague across the water, who has just solved something that I’m struggling with, I want instant access to that information.
Pharma IQ: Is that working effectively for you in GSK?
S Thomas: It wasn’t up until recently, we still had an awful lot of our data on paper for regulatory submission and there was an element of having to know where the answer was before you could go and find it.
Our colleagues in the US had used a certain brand of software for just storing their analytical data, but hadn’t used it to map metabolic outcomes. So, we took a look at that over the last couple of years and had a pilot trial, which allowed us to map data onto biotransformations. It wasn’t really designed for it, which had a chemist’s tool of A plus B goes to C plus D. But we took it and in a metabolic world, your parent can go up to 100 metabolites, so all through the alphabet four times over.
Pharma IQ:  So, are you storing your data more accurately now and had it become easier to access?
S Thomas: Far easier to access. Before the reports were printed out on paper; your chance of finding that piece of data that would help you in your day-to-day work quickly was slim. We needed to have known somebody who had seen the weird metabolites just by chance.
The ability using this data to search from almost any facet so you’re not limited to search by structure. Now I can search by mass, project, analyst or the site that it was done. I can ask for everything that was done in safety species, or say, that was done two years ago as it was seen in bile. We coined the term “data cube”; you can turn the face of the cube what you’re interested in and drill down from whichever piece of information you’ve got.
Pharma IQ:  What does the future looks like for GSK around your informatics?
S Thomas: There are on-going conversations with the manufacturers of the software, where we’ve got on-going input into the future direction. They’ve got some great programmers, but if you’ve got programmers in isolation, they might build something because they can and the chemist is sitting in the lab wondering what’s possible; getting these two people talking is what gets the software to be aligned with our work processes and makes things happen.
There are hopefully plans for the future that we can extend from searching our own data to be able to have internet searches on the same substructures. Say, you’ve got 18 or so hits within your internal database, but there are also 400 out on the web, you’ve got your own data internally, which is your quality, and added on to that, you can also go outside and search the quantity available from the internet.
Written by Niamh Madigan
Have Your Say
Rate this feature and give us your feedback in the comments section below
Please note that we do all we can to ensure accuracy within the translation to word of audio interviews but that errors may still understandably occur in some cases. If you believe that a serious inaccuracy has been made within the text, please contact +44 (0) 207 368 9425 or email helen.winsor@iqpc.co.uk