How AI Drug Discovery Technology is Transforming the Biopharma Industry

The expected return on investment from drug development is drastically falling.

While in 2010, returns averaged at 10.1%, in 2019 ROI dropped to just 1.9%.

Drug development typically costs upwards of £1 billion and can take anywhere between 8-12 years. Drug discovery is a key contributor to that time frame, and usually consumes a third of the total R&D budget. 

The future of the industry relies on finding new, innovative ways to speed up discovery and development without compromising the quality and accuracy of the process, and AI is proving itself capable of doing just that. 

With AI, biopharma organisations can improve the accuracy of predictions at the discovery stage, curbing the cost of the overall drug discovery process, and delivering better drugs to market, faster.

Read on to find out:

• How AI improves drug discovery
• How leading pharmaceutical organisations are using AI
• What the future holds for AI in medicine

How can AI improve drug discovery?

Less than 10% of drug candidates that make it to clinical trials progress to market.

A lack of knowledge on the 3d structure of drug compounds and targets, as well as a reliance on animal models that fail to accurately represent human physiology, are just two factors underlying this limited success rate.

Jackie Hunter, CEO of Benevolent Bio, a division of Benevolent AI, says that half of those failures are fundamentally due to a lack of efficacy. In Ernst & Young’s biotech report, she states, “that tells you we’re not picking the right targets”, and that “even a 5 or 10% reduction in efficacy failure would be amazing”.

AI has a vital role to play in clinical research.

By leveraging data and machine learning, AI technology can improve efficiency across every stage of the drug discovery process, including:

• Initial prediction of the target protein’s role in disease
• Design of in silico compounds from libraries to improve structural knowledge
• Novel target identification 
• Prediction of the structure – activity relationship
• Prediction of ADMET properties
• Selection of ideal patient population for clinical trials
• Observation of adverse effects

Recent pioneers in AI and Biopharma

Exscientia

Excscientia’s ‘Centaur Chemist’ is showing huge potential, and results for the pharmaceutical industry. ‘Centaur Chemist’ works by computationally sifting through and comparing millions of small molecules, choosing the most viable candidates for synthesis, testing and optimisation, before finally selecting the best drug candidate for clinical trials. 

Working with GSK, Exscienta successfully identified the first active lead molecule for the treatment of chronic obstructive pulmonary disease (COPD) – paving the route for a vitally needed novel therapy. 

Sanofi has also chosen to work with Exscientia, and have invested approximately $5.2 billion in aggregate into their collaboration. Their research will focus on 15 novel small molecule candidates across oncology and immunology.

Meanwhile, Exscientia has been contributing to the COVID-19 effort, and have announced a new project optimising novel small molecule inhibitors into a therapeutic development candidate for SARS-CoV-2 and other coronaviruses.

Recursion

Recursion is a digital biology company with AI at the centre of their work.

Recursion’s 4.7-petabyte biological image database fuels an advanced AI platform that can reveal drug candidates, mechanisms of action, novel chemistry and potential toxicity.

Just 18 months after their collaboration with Takeda in 2017, Recursion delivered new therapeutic candidates for more than 6 diseases. In 2020, Recursion identified TAK-733, a clinical-stage MEK inhibitor with the potential to treat hereditary cancer.

Chris Gibson, co-founder and CEO of Recursion announced “TAK-733 is a great example of the power of our approach to decode challenging and important areas of biology.” He describes the power of Recursion’s approach “by applying machine learning to images of cells, we capture cellular changes accompanying hundreds of unique biological perturbations, and even loss of just a single gene”.

Benevolent AI

Benevolent AI is a leading clinical-stage AI drug discovery organisation, that started working with AstraZeneca in 2019 to identify new targets for chronic kidney disease, and idiopathic pulmonary fibrosis. 

Since then, Benevolent AI’s Platform has identified two novel targets that have been experimentally validated and selected for AstraZeneca’s portfolio.

Professor Maria Belvisi, SVP and Head of Research and Early Development at AstraZeneca spoke about the collaboration positively, saying “our ongoing collaboration with Benevolent AI has enabled us to leverage the world’s available scientific literature and our in-house experiments, all brought together through machine learning to identify previously unrecognised links.” 

What does the future hold for AI in medicine?

While AI use in pharma has, until recently, focused on analysing large data sets, there remains a future potential to combine insights from real-time target population interactions with larger datasets to drive the design of personalised precision drugs.

According to Frank Nestle, Global Head of Research and Chief Scientific Officer at Sanofi, the “application of sophisticated AI and machine learning methods will not only shorten drug discovery timelines, but will also help to design higher quality and better targeted medicines for patients”.

Personalised precision drugs have long been conceptualised in medicine but remain too expensive to manufacture without AI intervention. In the future, personalised drugs could provide specific treatment to individual patients, based on their personal genetic background and symptoms to avoid side effects. This would aid early intervention and improve treatment adherence.

Before this level of disruption and development is achieved, the biopharma industry will need to pave the way with the right AI professionals and technical experts to ensure the right decisions are made at the right time.

Skills gap in AI

Ian Marison, founder and Chief Executive Officer of the Biofactory Competence Center (BCC) in Fribourg, Switzerland has written on the skills gap subject.

According to Marison, “a survey last year by the Coalition of State Bioscience Institutes found that traditional manufacturing positions were the easiest functional roles to fill. But when you look at the skill sets required to manage biopharmaceutical manufacturing processes — in particular, around engineering, data analytics and process development — the skill set shortage challenge persists.”

He notes that the use of artificial intelligence (AI) and process analytical technology (PAT) in drug production, as well as various approaches using data science, big data, and machine learning, is increasing the demand for life scientists with mathematical and computing skills. 

Are you facing talent shortages in AI?

Our Search by SRG Team are experts in STEM recruitment and can help you find expert talent to meet your business requirements. Email solutions@srgtalent.com and one of our consultants will be in touch to arrange a call at a time that suits you.

About the author: Janne Bates specialises in resourcing top talent within AI, Machine Learning and Bioinformatics across the Life Science industry. She partners with companies across the globe with a focus on the European market.

Specialist Areas: AI

The typical roles Janne recruits include from C suite down: Bioinformatics, Computational Biology, Computational Chemistry, Machine Learning, Data Science, Biological Systems Specialists