The quest for new medicines is a high-stakes gamble, with the odds heavily against success. The pharmaceutical industry pours billions into a process fraught with uncertainty, where over 90% of drug candidates fail to make it to market. Each failure means not just lost revenue but broken hopes for patients in desperate need.
The threat of emerging global health crises, including the looming spectre of Disease X — an unknown pathogen that could spark the next pandemic — also highlights the urgent need to accelerate drug discovery.
Drug discovery is a complex, interdisciplinary process that blends chemistry, biology and computational science. It begins with identifying a biological target — a molecule or pathway associated with a disease — before scientists search for compounds that interact with this target in a way that could lead to a new treatment.
The problem is scale. The so-called “chemical space” (or the theoretical universe of all possible molecules) contains an estimated 10,180 potential compounds. Scientists have barely scratched the surface, testing only a tiny fraction for potential use in medicine.
To speed things up, pharmaceutical companies use massive compound libraries, systematically screening molecules for biological activity. But even these databases are too vast for humans to sift through. That is where high-throughput screening comes in: robots testing thousands of compounds at breakneck speed, hunting for promising leads.
“Technological advancements, particularly in artificial intelligence (AI) and automation, are revolutionising drug discovery. AI is used to expedite drug identification, streamline design, and hence can drastically reduce costs and time to market,” says Linda Zhang, vice-president and head of APeC Commercial, Science and Lab Solutions at Merck Life Science.
See also: Singtel partners with Palo Alto Networks to bolster enterprise security
Revolutionising drug discovery
Merck Life Science is among the firms helping to revolutionise an industry notorious for its slow and costly innovation by using AI.
“Merck sparks discovery with AI and for AI. Our goal is to enhance efficiency and productivity in the drug discovery and development process for our customers who are the scientists in the lab,” Zhang says, adding that AI is integrated into multiple stages of drug development, from identifying promising molecular software to optimising synthesis routes.
See also: StarHub leverages Red Hat and F5 for its Cloud Infinity
For instance, catalysis is a crucial step in drug development, accelerating chemical reactions to make the process faster, cheaper, and more precise. But with hundreds of potential catalysts available — and each chemical reaction requiring a unique approach —finding the right one can be time-consuming and resource-intensive.
Merck’s Catalexis helps accelerate this process. The data is input into an AI-powered algorithm, which in turn generates a suggestion of which catalyst is optimal for the specific catalysis more quickly and by using fewer resources.
Meanwhile, Merck’s AI-driven drug discovery software Synthia is helping pharmaceutical companies streamline research and cut costs. “Synthia can provide candidate synthesis methods and routes, as well as promising solutions to dramatically reduce research time and costs,” says Zhang.
More than half of the world’s top 30 pharmaceutical firms have adopted Synthia, including GlaxoSmithKline (GSK), which uses it to enhance brainstorming and synthesis in chemical development.
In South Korea, JW Pharmaceutical and Daewoong are leveraging Synthia to automate synthetic research, accelerating the discovery, verification, and monitoring of potential drug candidates.
Merck also offers Aiddison, which combines generative AI, machine learning and computational drug design to boost the success rate of drug discovery and therapeutics.
This AI-powered software-as-a-service platform bridges the gap between drug design and synthesis feasibility by integrating directly with Synthia’s application interface. ISO 27001 certified, Aiddison enhances efficiency and precision in the drug development process.
To stay ahead of the latest tech trends, click here for DigitalEdge Section
Zhang says: “Based on experimentally validated data sets from Merck Healthcare’s drug discovery process, Aiddison accelerates the drug discovery process by identifying drugs with the right potency, solubility, safety and non-toxicity from more than 60 billion compounds and suggesting optimal synthetic routes.”
As for Merck’s mPredict, it offers rapid formulation solutions for faster drug formulation and quicker patient access. The AI-based tool can predict potential compatible co-formers, including compounds that may have been previously overlooked. This significantly shortens the formulation process and accelerates the successful delivery of new medicines to patients.
Role of humans and ethics
Despite AI’s growing role in drug discovery, Zhang asserts that human expertise remains indispensable. While AI can provide predictions based on the data available, the results must then be validated and interpreted by human researchers. “AI-based approaches are not a substitute for traditional experimental methods, and AI cannot replace the expertise and experience of human researchers.”
Drug discovery is riddled with uncertainty, requiring judgment calls based on incomplete data.
“Experienced researchers can weigh factors such as biological plausibility, clinical feasibility, and potential for success, which requires expertise that goes beyond what AI can currently achieve. While AI enhances the drug discovery process by accelerating data analysis, pattern recognition and prediction, human expertise is essential for generating hypotheses, making critical decisions, ensuring ethical standards, and navigating the complex biological and regulatory landscape,” says Zhang.
Recognising the ethical impact AI has in our society, Merck has developed a Code of Digital Ethics that defines its principles in handling data and AI. It is based on five core values: justice, autonomy, beneficence, non-maleficence and transparency.
“These five principles help us to create specific guidelines for certain applications or develop products responsibly. We are convinced that algorithmic systems should be explainable, and the users of our digital services should be aware of the direct and indirect effects created by automated decisions,” adds Zhang.
Backing this commitment, Merck launched a digital ethics advisory panel in 2021 to oversee responsible AI use.
The panel, notes Zhang, plays a crucial role in ensuring the company develops new digital technologies responsibly and addresses potential ethical issues arising from the use of those technologies. It consists of renowned academic and industry experts in the fields of digital ethics, legal and regulatory, big data technology, digital health, medicine and data governance from the US and Europe.
Looking ahead
To further help accelerate drug discovery, Merck’s research teams are actively exploring and developing AI algorithms and contributing to initiatives like the Machine Learning Ledger Orchestration for Drug Discovery (MELLODDY), a consortium focused on collaborative AI models for pharmaceutical research.
Since open innovation is a key part of its strategy, Merck taps on external expertise through partnerships with AI start-ups and research organisations. The company has launched initiatives, such as the Synthia Compound Challenge, a competition to identify the best submitted synthetic route for a given small molecule.
It also collaborates with the Acceleration Consortium at the University of Toronto to make an AI-driven experimentation planner, Bayesian Back End (BayBE),available open-source on GitHub.
“At Merck, we are committed to staying at the forefront of innovation in AI-driven drug discovery. We invest in leading-edge technologies and collaborations with academic institutions, research organisations, and technology companies. Also, by engaging in open innovation initiatives, we can tap into a broader range of expertise and ideas, which helps us stay agile and adaptive in a rapidly evolving landscape,” concludes Zhang.
