What could artificial intelligence mean for pharma?
pharmafile | January 15, 2016 | Feature | Medical Communications | artificial intelligence, digital, digital pharma, medical technology, medtech, tech
With ever rising costs in research and development, the frustratingly long time spent in bringing new novel drugs to market and the high rate of failure in these processes needs to be tackled.
The majority of drugs take a decade or more to come to market, cost billions, and can even ruin a company if they fail in late-stage trials having poured in so much investment.
Step forward Artificial Intelligence (AI); a concept becoming more and more important in addressing these issues and looking increasingly like the future of drug discovery.
Faster and better drug discovery
Perhaps the most obvious application of artificial intelligence in pharma is using its ability to quickly ‘read’ vast amounts of scientific data: research published in journals, as well as patient records and tissue/blood samples, and using patterns in the data to make scientific hypotheses which can direct pharma companies’ drug development.
The speed of AI in these processes allows companies to develop drugs based on biological markers, with greater accuracy, rather than the scattergun approach of chemical screening. In this way, companies can zero in on particular indications which the drug is most likely to successfully treat.
Boston-based biotech Berg’s Niven Narain says the company’s platform, Interrogative Biology, allows it to look at 14 trillion data points in a single tissue sample. He ambitiously claims that artificial intelligence will halve the time (and potentially the cost) it will take to bring its lead candidate BPM31510 to market.
Similarly, IBM’s Watson supercomputer is currently involved in trials where it scans mutation data from the tumours of 20 brain cancer patients. This is something that would take human scientists weeks and months to analyse, but Watson can do the same in a matter of minutes.
Through machine learning, Watson and similar platforms get better and faster at the process. Ultimately, the screening process could be fast enough to analyse the entire genome of each patient’s individual cancer and for treatments to be tailored based on its specific mutations, if they exist. If not, there will be a company interested in putting that right.
In the UK, the University of Manchester’s AI platform, known as Eve, can screen more than 10,000 compounds in a day, matching them to likely targets. Again, through machine learning and hypothesis testing, Eve recognises why ‘she’ has succeeded, and so gets faster the more screening she performs.
Lower drug pricing
Cheaper drug development should enable cheaper prices. Drug pricing is a hugely controversial issue in the industry nowadays, and the reputations of pharma companies are suffering as a result of massive hikes. Pharma bosses will often justify such increases by citing the huge costs of R&D, so if such costs can be significantly reduced – as Narain suggests is possible – they will no longer be able to use this justification, and prices should (in theory) fall.
Beyond discovery
It isn’t just in the discovery process that AI can help pharma companies, but right up to approval and even in the general running of the companies. After a promising candidate is discovered, AI could be used to design more effective clinical trials and more quickly analyse the data that emerges from them.
Even business decisions may be handed over to supercomputers. Consider the huge numbers of mergers and acquisitions last year. AI could more effectively analyse potential synergies gained from the merger of particular companies, allowing them to decide if a combination is worthwhile. If it is, then AIs can help make decisions on integrating the R&D departments, for example.
It could also have a hand in the sales and marketing process. Last year, Eularis released a cloud-based marketing analytics platform for the pharma industry, backed by cutting edge algorithms and the same machine learning ability used in Google’s driverless cars. Named E-VAI, it has the ability to learn from the successes of marketing campaigns in the last decade and effectively mimic them and apply them to new products.
The application of AI in pharma is in its infancy, and it could take two decades to reach its full potential. However, the beginnings of a technical revolution that could change the way in which drugs are brought to market appears to have begun, which is good news for pharma companies and patients alike. Where there is data to be analysed or a business decision to be made, the betting is that the AIs of the future will challenge any current pharma executive to do it better and faster.
Joel Levy
Related Content
Virtual Inspections: Embracing the Digital Change
Virtual inspections must become a part of companies’ repertoires if they aim to maintain quality …
gaitQ awarded NIHR grant for Parkinson’s wearable tech
The UK’s National Institute for Health Research (NIHR) has awarded a £135,000 grant to gaitQ, …
Microsoft and UCB collaborate to treat immunological and neurological diseases
UCB and Microsoft have agreed a new multi-year collaboration to discover new medicines by combining …
