As AI-designed drugs start to enter large-scale clinical trials, DDW’s Diana Spencer investigates how new digital tools are reinventing and reshaping drug discovery for the future.
From finding potential active substances and identifying novel targets, to simulating how drugs will function in the human body and optimising laboratory workflows and informatics, digital technologies are having a significant impact on drug discovery and development. Technologies like artificial intelligence (AI), machine learning (ML) and robotics/automation are enabling researchers to process enormous data sets and conduct preclinical investigations in record time, speeding up the process and offering greater accuracy.
However, many have cautioned against too great a reliance on these tools and emphasised the importance of checks and balances to ensure the systems are properly regulated and results are verified.
Companies leading the way
There are a number of companies leading the charge in applying innovative technologies, particularly artificial intelligence (AI) and machine learning (ML), to the search for new medicines.
Seven AI-focused companies were highlighted by Clarivate in its 2023 ‘Companies to Watch Report’1:
- AQEMIA, which uses quantum and statistical mechanics algorithms to fuel generative AI to design novel drug candidates.
- Auransa, which harnesses available data to discover highly intractable diseases with poorly understood biology.
- Enveda Biosciences, which leverages a combination of ML, metabolomics and automation to identify bioactive, plant-based molecules of interest.
- Pharos iBio, which utilises AI technology to develop treatments for refractory and rare diseases such as cancers and repurpose existing therapies.
- Quris-AI, which combines patient-on-a-chip, AI and real-time nano-sensing technologies to predict the safety and efficacy of drug candidates.
- Relation Therapeutics, which uses high-resolution biology, ML and clinical insights to discover transformational medicines.
- SandboxAQ, which combines AI and quantum technology (AQ) to advance drug discovery for complex, undruggable targets.
Companies like these, that can offer digital technology insights are – unsurprisingly – enjoying considerable popularity. Biopharma and biotech companies without these capabilities are keen to harness these technologies and the number of strategic partnerships with such companies is accelerating at an unprecedented rate. Sanofi has stated that it intends to become the first pharma company “powered by AI at scale”, and as part of this ambition, agreed a collaboration with BioMap to co-develop advanced AI models and protein Large Language Models that that it hopes will enable biologics design and multiparametric optimisation2.
In 2024, Amgen announced plans to use NVIDIA’s DGX SuperPOD-powered insights from human datasets to build a human diversity atlas for drug target and biomarker discovery. VantAI and Blueprint Medicines also extended their existing alliance to focus on ‘undruggable’ targets using induced proximity drug discovery.
In another early 2024 deal, one-to-watch SandboxAQ acquired Good Chemistry, a computational chemistry company that uses AI, quantum and other advanced technologies to accelerate medicines research3.
Good Chemistry founder and CEO Arman Zaribafiyan comments on the importance of these new high-tech resources: “The rapid advances in AI, quantum, cloud and high-performance computing have unlocked endless opportunities for companies like Good Chemistry and SandboxAQ to reinvent the way we think about chemistry, discover ways to make products safer, stronger and more sustainable, and reshape the fabric of our world.”
Universities are also using these digital technologies to advance their research. Late in 2023, SimBioSys and the Chan Lab at UT Southwestern revealed plans to enhance PhenoScope’s tumour bank with spatial transcriptomic data with the aim of developing cancer biomarkers for immunotherapy.
Accelerating drug discovery
In turn, this support and investment in new technology is having a meaningful impact on the speed and efficiency of drug discovery and development.
One hurdle in drug discovery is finding new chemical structures that have drug-like properties, and this is an area where AI is coming to the rescue. Biotech company Gero used a hybrid quantum-classical machine-learning model to interface between classical and quantum computational devices to generate novel chemical structures for potential drugs. It was successful – the model suggested 2,331 novel chemical structures with properties typical for biologically active compounds, less than 1% of which had a high similarity to any molecule in the training set4.
“These breakthroughs pave the way for a dramatic acceleration of the drug discovery process,” says Peter Fedichev, CEO of Gero. “Drug design operates at the intersection of the realms of classical and quantum phenomena, and requires simultaneous determination of quantum properties of drug-like molecules and their effects on living systems described by classical physics. This is why quantum computing will significantly augment our capacity to develop transformative treatments for the most challenging diseases and conditions, including ageing itself.”
Digital tech is also helping researchers to find target candidates for some of the most challenging disease areas. Researchers from the University of Oslo, University of Chicago Pritzker School of Medicine, and Insilico Medicine used an AI target discovery engine to analyse transcriptomic data to identify dual targets for cancer and ageing5. The research identified a number of potential age-associated cancer targets and the team was able to validate one of most promising candidates, the gene histone demethylase or KDM1A.
“We were very encouraged by the findings,” enthuses Evgeny Izumchenko, Assistant Professor of Medicine at UChicago in the section of haematology and oncology. “This is a first study showing the feasibility of AI-driven approaches to identify potential dual-purpose targets for anti-ageing and anti-cancer treatment, and clearly demonstrates the value of such tools in addressing the complex challenges at the interface between ageing and carcinogenesis.”
We are starting to see the visible impact of new technologies on the sector, as drugs discovered through the use of AI start to enter clinical investigations. Verge Genomics’ potential amyotrophic lateral sclerosis (ALS) treatment VRG50635 is one of the first drugs to enter clinical trials that was entirely discovered and developed using an AI-enabled platform6. The company’s platform evaluated more than 11.4 million data points from ALS patient tissue and genetics datasets to discover loss of endolysosomal function as a new causative mechanism in ALS, and uncovered PIKfyve as a promising new therapeutic target. VRG50635 has so far improved survival in ALS patient neurons and has shown efficacy in multiple preclinical studies in ALS-relevant models of motor neuron degeneration.
Another AI-powered drug discovery and development company BPGbio has moved to Phase II development of its lead candidate for advanced pancreatic cancer treatment BPM31510. So far, the results show BPM31510 was well-tolerated, doubled progression free survival vs. chemotherapy alone, and have prompted further investigation of BPM31510 as a first-line therapy.
Taking on antimicrobial resistance
Antimicrobial resistance (AMR) or antibiotic resistance is one of the greatest threats to human health facing society today and digital technologies are playing a role here too.
One company using AI models to tackle the problem of AMR is Obulytix, a spin-off based on research results from Ghent University and KU Leuven7. The company has created a novel AI platform to develop enzymes from bacteria-killing viruses (bacteriophages) as a new way to tackle bacterial infections, and has secured €4 million investment to support its development.
Obulytix is not alone. In late 2023, Massachusetts Institute of Technology (MIT) researchers revealed that they had identified a new class of antibiotic candidates against gram-negative bacteria using deep learning8. The newly discovered compounds can kill methicillin-resistant Staphylococcus aureus (MRSA) grown in a lab dish and in two mouse models of MRSA infection. Studying the AI model itself also provided the team with additional insights into how to identify potential antibiotics. The researchers were able to figure out what kinds of information the deep-learning model was using to make its antibiotic potency predictions, which could help researchers to design additional drugs that might work even better.
“The insight here was that we could see what was being learned by the models to make their predictions that certain molecules would make for good antibiotics. Our work provides a framework that is time-efficient, resource-efficient, and mechanistically insightful, from a chemical-structure standpoint, in ways that we haven’t had to date,” says James Collins, the Termeer Professor of Medical Engineering and Science in MIT’s Institute for Medical Engineering and Science (IMES) and Department of Biological Engineering.
The robotic revolution
Once targets and molecules have been identified and simulations run, the next step in the process necessitates a move to the laboratory to validate the candidate. This is where robotics and automation are making a difference, often in conjunction with AI models.
In June 2023, lab automation company Opentrons launched its Flex robots for scientists using generative AI. “For far too long, scientists have been constrained by their laboratory tools,” explains Jon Brennan-Badal, CEO of Opentrons. “By making lab automation as easy as using a smartphone, the Flex robot democratises access to automation in life sciences research.”
Automation is also being applied to genomics sample preparation, an area that requires particular robustness and scalability. The Advanced Sequencing Facility at the Francis Crick Institute is working with company Automata to automate workflows, with the aim of decreasing manual touchpoints, increasing throughput and optimising R&D flexibility. The walkaway workflows are allowing the researchers at the facility to prepare samples and generate data rapidly, including the validation of CRISPR genome editing, the extraction of genetic material from tumour samples and genomic surveillance of the Covid-19 pandemic10.
Regulating digital tech
Despite the breakthroughs as a result of these new insights, some have urged caution, particularly in relation to a greater reliance on AI and ML. The European Medicines Agency (EMA) published draft guidelines on the use of AI in July 2023, highlighting that a human-centric approach should guide all development and deployment of such technologies12.
The EMA report notes that AI and ML tools could be used to replace the use of animal models in preclinical development, to support the selection of patients for clinical trials, or to draft, compile, translate, or review data. This breadth of applications brings challenges such as the understanding of possible biases in the algorithms, the risks of technical failures and the wider impact these would have on AI uptake in medicine development.
“The use of AI is rapidly developing in society and as regulators we see more and more applications in the field of medicines. AI brings exciting opportunities to generate new insights and improve processes. To embrace them fully, we will need to be prepared for the regulatory challenges presented by this quickly evolving ecosystem,” says Jesper Kjær, Director of the Data Analytics Centre at the Danish Medicines Agency and co-chair of the joint HMA-EMA Big Data Steering Group (BDSG).
Biosecurity is an issue that has come to fore more recently due to the development of these new technologies, especially ‘biodesign’ tools, specialised AI models that are trained on biological data and provide insight into biological systems. The Federation of American Scientists (FAS) has also published recommendations to address the need for oversight of biodesign AI tools, biosecurity screening of synthetic DNA and guidance on biosecurity practices for automated laboratories13.
The recommendations include institutional oversight for these tools from the government, the need to establish standards for evaluating their risks, expanded infrastructure for cloud-based computational resources, biosecurity screening of synthetic DNA, and government guidance on biosecurity practices for automated laboratories.
FAS comments: “AI is likely to yield tremendous advances in our basic understanding of biological systems, as well as significant benefits for health, agriculture, and the broader bioeconomy. However, AI tools, if misused or developed irresponsibly, can also pose risks to biosecurity. The landscape of biosecurity risks related to AI is complex and rapidly changing, and understanding the range of issues requires diverse perspectives and expertise.”
About the author
Diana Spencer is Senior Digital Content Editor on DDW. She brings 18 years’ experience in both print and digital editorial roles for B2B publications dedicated to the life sciences sector, including medical and pharmaceutical industry journals/magazines.