Artificial intelligence and big data are sweeping through the healthcare and pharma markets as cloud computing and low-cost monitoring devices enable patients’ reactions to be easily recorded and stored.
But what about the chemistry behind how the drugs are made?
Here, very little has changed and processes in place in 150 years ago remain the norm according to Mark Warne, the chief executive of AIM-listed DeepMatter PLC (LON:DMTR).
Chemists like to think of themselves of artists, he adds, using their years of experience to concoct compounds and molecules in exactly the right quantities and ambient conditions.
The problems arise when they try to pass on the methodology.
Any follower of recipe books will testify that a page of instructions leaves plenty of room for misinterpretation.
It is a similar story with drugs except it is a lot more expensive.
Most formulations are noted in books or journals and as many as 50% are impossible to reproduce again due to the articulation/interpretation problem.
It is an issue that costs the pharma industry in the US alone US$28bn a year.
DeepMatter’s combination of hardware, software and machine learning offers a solution.
Its system is aimed at creating a bridge between research and process chemistry, disciplines that often barely interact with each other.
The research team discovers a molecule and passes it up to the process team, which would then work out its own way to produce it.
It’s an inefficient approach but Warne believes he can engender change.
“We are digitising chemistry,” he says.
“Bringing machine learning and AI to productivity in chemistry is something unique.”
DigitalGlassware is the name of the platform.
Reactions are accurately and systematically recorded, translated into computer code and stored in a shared facility in the cloud.
Storing data in that way means scientists can collaborate and share details of experiments from anywhere and in real-time.
“Typically, chemists might not look at the inside of glass during an experiment but that information is available,” adds Warne.
And that helps you measure data from when the reaction takes place rather than at the end,
Warne foresees a step-change in productivity and drug discovery, less duplication and waste and a faster route to discoveries.
DeepMatter deployed the platform for the first time last year in a pioneer programme of seven global organisations comprising large pharma companies, life science reagent providers, contract research organisations (CROs) and academics.
The aim was to collect reaction data and funnel that back to the chemists to enable them to make the molecules in real-time while also honing the planning stages.
A key feature is that the DigitalGlassware system not only records what happens, but also what didn’t, giving critical insights not only into why a process worked but also why it might have failed.
The pioneer programme was designed to run for twelve-months and DeepMatter will wind it up at the end of the year.
One company, O2H Discovery - a Cambridge-based CRO, has now graduated from the programme to become the first paying customer and Warne is hopeful others will follow.
It’s a question of whether the commercial terms are acceptable.
Pharma, too, an industry is notorious for its slow pace of change.
Even so, given the sums involved and the costs currently, Warne is confident.
“It costs billions to make the molecules, but we will charge tens to hundreds of thousands of pounds to use the technology.”
DeepMatter listed on AIM just over a year ago and the shares have been steady performers subsequently.
This year’s figures will reflect the early stage of its development but once more contracts have been signed, the company will have a better view on how revenues will grow and will be able to put some numbers out, says Warne.
A chemist himself, he says the industry is ripe for modernisation and change is afoot he believes.
For example, historically, a report was written after the reaction detailing what happened, but increasingly the write-up is before the experiment and says what the chemist expects to happen.
That might not sound like much, but in the chemistry business it is a big deal and all only possible with the advent of cloud computing and low-cost sensor arrays.
“A ‘framework’ to transfer knowledge is the way the industry will go,” says Warne and DeepMatter is in the front rank of those leading it there.
Admittedly it is early days, but with a market cap of £21mln at 2.85p there looks to be lots of potential.