AI-driven drug discovery (2021)
The startups turbocharging the hunt for breakthroughs
Last updated: 2 Sep 2021
Market 101
Inventing new drugs is not only notoriously slow and expensive but has become harder over the decades — an observation industry wags refer to as Eroom’s Law (Moore’s Law spelled backwards). In search of a little elbow grease, pharmaceutical companies are turning to artificial intelligence for help combing through hundreds of millions of chemical compounds (with results promised in hours rather than months.
There’s just one problem: pharma players generally lack the expertise to develop these methods in-house. They’re up against a new competitor — or potential partner — in Big Tech, with startups ploughing millions of pounds into AI drug screening. Teaming up with these labs offers pharma players the combined brainpower of engineers and biologists — which might just be enough to weaken, if not fully invalidate Eroom’s Law,
The Basics Drug discovery is the search for the best possible match between two molecules: one that can be developed into a successful drug compound, and one that is associated with a disease in our bodies. To be effective, the drug compound’s structure needs to fit and bind to the target molecule.
Early stage market map
Key Facts
€1.7bn
Average cost of R&D per drug that makes it to market1
1/3
Amount of R&C cost going to drug discovery phase2
45%
Potential growth in pharma earnings with AI help3
Trends to watch
1. Growing independence
→ AI startups rely on incumbents as collaborators and customers, but many no longer wish to be acquired by them.
→ Increased VC funding may help startups maintain their independence and continue to develop their own drugs.
2. Cracking open data troves
→ Various policies aim to boost data sharing among medical researchers in Europe, such as the EU’s forthcoming science cloud initiative, which pledges to increase access to publicly funded research.
→ This will benefit startups, which need access to large datasets to train their AI systems.
3. New data, new opportunities
→ Genomic data would allow researchers to design drugs based on human biology from the outset, rather than relying on animal models.
→ Combining big datasets with smaller ones that focus on individuals within the target population could generate more precise insights.
4. Quantum: the next frontier
→ Quantum mechanics describe the behaviour of atoms and molecules at their most granular levels. Quantum calculations however require huge computing effort.
→ Researchers currently rely on simplified, less accurate representations of molecules. As computing power grows, expect to see the quantum world begin to take on drug discovery.
Startups tracked by Sifted
Sifted take
AI could make drug discovery faster, more accurate and more efficient. But there’s a difference between startups screening known chemical libraries and those generating new digital molecular structures to bind with a specific target. The latter approach may be more difficult, but potentially more groundbreaking — so keep an eye out for any breakthroughs.
Rising stars
Sorts patients into clinical trials using blood-based screening. The company’s tech promises to achieve better trial outcomes while saving on costs.
Round
Seed
Valuation
€13.5m
Date
2019
Size
€1.5m
Developed AI to generate insights from very small datasets. Founder Noor Shaker previously ran Kuano, another AI drug discovery lab that has raised €2.5m to date.
Round
Grant
Date
Size
€517k
Combines a quantum-inspired algorithm with AI to accurately predict the affinity between a drug compound and its target. It’s able to generate novel molecules for drug compounds based on physics rather than datasets. Cofounders Maximilien Levesque and Emmanuelle Martiano Rolland are a former quantum mechanics researcher and management consultant, respectively.
Round
Seed
Valuation
€8m
Date
2019
Size
Early stage startups to watch
Antiverse
Identification of new antibody therapeutics
€2.1m
€1.7m
-
Aqemia
De novo drug design
-
-
-
Arctoris
Identification of new drug targets
€9m
€7m
-
Genome Biologics
Identification of new drug targets
€2.5m
€2.4m
-
GlamorousAI
Identification of new drug candidates
€528k
€517k
-
ILoF
Clinical trials
€4.5m
€1.5m
-
Kuano
Identification of new drug candidates
€3.7m
€1.2m
-
MAbSilico
Identification of new antibody therapeutics
-
-
-
Molecule.one
Identification of new drug candidates
€4.5m
€3.9m
-
Novai
Identification of new drug candidates
€1.3m
€800k
-
SOM Biotech
Health
€34m
€7m
-
Europe’s success stories
Who early stage startups are up against
(Pre-)Seed
Series A
Series B
Series C
Series D+
IPO/Exit
→ Has pinpointed potential arthritis drug baricitinib, which is now part of a clinical trial
→ Has ongoing partnerships with pharma giants Novartis and AstraZeneca
(Pre-)Seed
Series A
Series B
Series C
Series D+
IPO/Exit
→ The Oxford-based company is the first to put an AIdiscovered drug into human trial
→ Has partnerships with drug heavyweights including Sanofi, Roche and GSK
Sources
Research reports
Digital innovation in the pharmaceuticals and chemicals industries | January 2021 | MIT Technology Review
1,2,4 Intelligent drug discovery, powered by AI | November 2019 | Deloitte
Measuring the Return from Pharmaceutical Innovation | 2019 | Deloitte
Market research
Drug Discovery Services Market - Global Forecast to 2025 | September 2020 | Markets and Markets
News articles
The Post-Covid Future of AI for Drug Development | June 2021 | Labio Tech
3 AI in Drug Discovery Starts to Live Up to the Hype | April 2021 | GEN
Roadmap: Unlocking machine learning for drug discovery | March 2021 | Bessemer Venture Partners
AI and Drug Discovery: Attacking the Right Problems | March 2021 | Science Magazine
Your feedback
How would you rate this briefing?