around Artificial Intelligence’s (AI) potential to transform the drug discovery
process has been building up in the biopharma industry. AI-drug discovery
starts-ups have boomed in the last decade and billions of dollars have already
been poured on the promise of an efficient and agile drug discovery process.
influenced all sectors of the drug discovery value chain. Technologies have
come up to analyze numerous medical images, to aggregate information from
millions of publications for rapid analytics, to design and optimize new drugs,
to design better studies, and recruit for clinical trials.
The biopharma companies
have been keenly watching the space and making definitive, but careful investments/partnerships
with a handful of AI start-ups. Over the past decade, more than 50 pharma
companies have signed 100+ partnerships for pilot programs with AI start-ups,
with more than 75% of these deals coming in the last 3 years. The early adopters are already
observing a positive trend in their R&D productivity.
2019 was a big
year, marking several AI milestones in drug discovery. Deep genomics identified
a new genetic target for Wilson disease, and subsequently a new oligonucleotide
therapy DG12P1. The whole process took less than 18 months, as against to 4-6
years in a “traditional” discovery process. UK-based AI start-up, Exscientia also
announced the delivery of a selective and potent in vivo lead molecule to
partner GSK. Later in the year, InSilico Medicine designed, synthesized, and
validated a new drug against DDR1, a well-characterized target for Fibrosis, in
mere 46 days. The time taken is approximately 15-fold lesser than the traditional
expectations soared early this year when Sumitomo Dainippon Pharma announced the
initiation of phase I clinical study of DSP-1181, a drug developed using AI, for
obsessive-compulsive disorder. DSP-1181 was created jointly with Exscientia and
required less than 12 months to reach the clinic, a fraction of the typical
average of 4.5 years’ timeline.
these early partnerships, a few leading pharmaceutical giants, like Pfizer,
Merck, Novartis, GSK, Roche, and Merck KGaA have taken giant leaps to adopt AI internally
as well. Some have created dedicated executive positions like Chief Digital Officer
or Head of AI to take steps towards building internal AI capabilities and
aligning their research workflows to start embracing the full potential of AI
The state of AI drug discovery in
Touting AI as
the next big disruptor, most countries, including the large Asian economies Japan,
China, and India have come up with guidelines and national strategies to build
AI leadership, and the pharmaceutical industry has been a key focus. The Asia Pacific AI market was
estimated at around the US $450 million in 2017 and is expected to grow at a CAGR
of 46.9% by 2022. Asia is likely to surpass the US as a hub for AI specialists,
as China and India are headed strong towards developing internal capabilities
in AI by nurturing the specialized start-up ecosystem.
While Asia has
developed a strong AI capability in many domains like finance, voice, and image
recognition, etc., the Asian AI drug discovery ecosystem is still in its
infancy. Asia accounts for only 5% of AI drug discovery companies globally,
China accounting for 3% alone. Also, unlike western pharma companies, only a
handful of large Asian pharma’s have only begun now to superficially tap the
potential of AI.
China has been
at the epicenter of the growth strategy for most global pharmaceutical
companies in the last few years, being the second biggest pharmaceutical market
after the US. In 2017, China published its ‘AI strategic plan’ to gain AI
leadership by 2030 and has been aggressive about it ever since. China is already considered the
leader in image recognition using AI, a technology transferable to some aspects
of drug discovery.
Pharmaceutical companies have not been very transparent on the adoption of AI.
However, the $1.5billion deal, one of the largest AI drug-discovery deals yet,
between Atomwise and Jiangsu Hansoh Pharmaceutical group—the Chinese company
behind 2019’s largest biopharma initial public offering (IPO), is indicative of
the momentum. While the adoption began very late as compared to developed markets,
it has found its way in aggressively in the last 2-3 years and deeper
penetration is expected in the next 5 years.
Medicine, recognized as one of the top 10 AI drug discovery companies, moved
its headquarters from the US to Hong Kong in April 2019, citing proximity to
tap enormous potential, availability of talent pool, and benefits in China. Within
a year, it has publicly announced partnerships with at-least 3 Chinese
pharma companies were among the first to explore AI with partnerships,
reflecting the deep-rooted culture of innovation. About 10 Japanese pharma
companies have at least one AI exploratory partnership. However, except for
only a few Japanese pharma companies like Sumitomo Dainippon, Takeda, Astellas participating
in consortiums to drive AI-based innovation, the adoption has been limited to
one exploratory partnership in the Chemistry based applications or repurposing.
The adoption of AI at scale and biology-based applications are still to find
their way into Japanese pharma.
National Strategy for AI, drafted by Niti Ayog in 2018, aims to position India
as the AI garage for 40% of the world, and healthcare/pharma is one of the 5
key sectors covered under it. Even though the govt. initiatives are based on a strong
premise, the translation has been sporadic.
of AI in Indian pharma has been limited. The only publicly available partnership
in AI drug discovery is between Sun pharma advanced research center (SPARC) and
Schrodinger. The lack of an innovation mindset is a main reason that Indian
pharma companies are still ‘naïve’ in terms of the adoption of AI for drug
Time to up the game
in the industry believe that companies that fail to adopt AI, may never catch
up. While a solid foundation in AI is being built, the broader adoption of AI
in Asian pharma is still to gain momentum.
companies have traditionally followed the west. The question is, now that Asia is headed to
become a hotspot for AI activity, will the Asian pharma companies will rise to
the occasion? Will they tap the potential of AI broadly or will they still be
apprehensive about a wider adoption and just follow the tail on where the west
is headed to?
leaders throughout Asia should prepare for a future where AI dominates the
business, where AI changes the nature of human work, where AI changes the way
we discover drugs, creating new opportunities and disrupting current workflows
Considerations for adopting AI
The key to
successfully adopt AI is to explore early and often. It is prudent to build
internal capabilities and realize what AI can and cannot do. Understanding the
quality and quantity of data required for each application is critical to
A safe bet to
begin would be to replicate what has been done and scale from there. The majority
of western pharma companies began their AI journey by exploring chemistry-based
solutions and using AI to generate patient-centric insights from real-world
data. A lot of
successful case studies of AI in drug discovery are now available.
scenario would be to adopt or partner with early-stage companies. Start-ups
would gain subject matter expertise and AI ‘naïve’ pharma companies will be
able to begin the journey at minimal early investments. However, it will be
important to filter out the noise and adopt platforms with solid AI engineering
and a proof of concept present.