GSK's chief data officer reveals how machine learning has been used to speed up drug discovery

Machine learning is driving a step change in the production of new medicines, says Dr Mark Ramsay

Data and analytics have become key topics for many boardrooms, right up there with shareholder relations and risk management. In data-driven businesses like pharmaceuticals this has been the case for some time, but other sectors are catching up. Dr Mark Ramsay, CDO, SVP and head of R&D at pharmaceutical giant GSK, sees this process as inevitable and universal.

"Some of that is what I refer to as the 'Amazon effect'," he said.

"You look at retail and how data and analytics was used to really transform that industry, and pretty much every sector is now recognising that that same type of transformation will occur. It's not a matter of if, it's a matter of when."

Within pharmaceuticals the focus of analytics has changed. Whereas two or three years ago it could be characterised as 'vertical'- looking at a very specific business function to a great depth - it is now moving in a horizontal direction, Ramsay said.

"We are now using data from one part of the R&D organisation to influence decisions in other parts of R&D. That's one of the big changes."

The other big change, he went on, is the use of machine learning and deep learning algorithms to automate more decisions in the research field. In particular, these methods are being used in drug discovery, the process by which new candidate medications are discovered.

Drug discovery is a lengthy process. It takes between 10 to 15 years from start to finish and costs on average $1.6bn to bring a new medicine to market. A lot of effort is therefore being expended on radically reducing these numbers, and much of that involves data and analytics.

In recent years there has been a huge increase in the amount of genetic material made available for analysis. GSK works in partnership with the UK Biobank giving it access to the DNA of 500,000 medical volunteers. By sequencing that DNA the company is able to increase significantly the quality and speed of the selection process for candidate medicines.

We are really targeting a step change not an incremental change with analytics now. It's transformational

"Typically it takes between six and eight years from an early discovery of a potential new medicine to the time it goes into clinical trials and our goal is to reduce that to less than two years by 2020," Ramsay said.

"Cutting out that time gets new medicines to patients faster which has a huge impact, so we are really targeting a step change not an incremental change with analytics now. It's transformational."

The changing role of the CDO

Ramsay is a founding board member of the International Society of Chief Data Officers, which was started by the MIT. As such he is well placed to comment on changes in the CDO role.

"It was started by Yahoo about eight or 10 years ago," he said.

"There was a fairly slow adoption, but it accelerated from there. It is growing quite rapidly but the market is kind of confused because now we have chief data officers and chief digital officers but it's all just about data and analytics."

Different organisations have chosen to adopt the role in a variety of ways, he went on, but "at the end of the day it's about how does the company maximise the value of data as a strategic asset".

That said, there are two distinct types of CDO - tactical and strategic. The latter, Ramsay believes may be in decline. They play a defensive role typified by a focus on data governance and data quality. This type of CDO role is most common in banks and other financial organisations that needed to clean up their act following the financial crash of 2008.

By contrast "strategic" CDOs are running forward at full pelt. They are key members of the board, often reporting directly to the CEO or president, and keenly focused on how to expand the business using data.

"The strategic CDOs are playing offense [as in the American football position], they're always looking at how to drive a lot more value into the organisation," Ramsay said.