(Last updated: July 2023)

The type of artificial intelligence (AI) which scares business leaders, experts, and activists all over the world, is called “general artificial intelligence” — the one which could “think” pretty much like humans do, and which could quickly evolve into a dangerous “superintelligence”. There is a notion that it might be invented in the nearest decades, but today we are definitely not there yet. However, with the recent groundbreaking advances in deep learning and natural language processing technologies, particularly — large language models (LLMs), we have all felt that the world might indeed be changing more rapidly than AI deniers used to think. Let’s face it, only few could foresee such an alarmingly efficient public release of the most generalized AI model of all time — ChatGPT, by OpenAI. Adding more to that, a race of LLMs has begun, with Google launching Bard, and other companies following the path. 

Even notoriously slow for technology adoption, the pharmaceutical industry has seen accelerated integration of various AI technologies over the last decade, and the interest is rapidly growing. The potential impacts of this transformation extend beyond healthcare providers and patients grappling with difficult-to-treat ailments, reaching into the biotech sector as well. Based on projections from Morgan Stanley Research, even slight enhancements in early-stage drug development success rates, facilitated by artificial intelligence and machine learning, might result in an additional 50 innovative treatments over the next decade. This could equate to a market opportunity exceeding $50 billion. 

According to a 2022 report by GlobalData, 50% of professionals within the healthcare industry would prioritize investments into AI, over other emerging technologies, such as big data (38%), digital media (37%), cloud computing (31%), real world evidence, RWE (27%) and others. 

The 2022 thematic research report titled ‘Artificial Intelligence (AI) in Drug Discovery‘ from GlobalData predicts that the total expenditure on AI by the pharmaceutical sector is projected to escalate to more than $3 billion by 2025.

Let’s review specific examples of how AI is used in the pharmaceutical industry. 

(Since most AI-driven companies use a mix of different approaches and rely on interdisciplinary sources of data for their modeling work, the below classification of AI use cases is illustrative):

AI for drug target discovery and disease modeling

One of the most promising areas of AI in pharma is modeling biological systems, and identifying novel drug targets. A number of AI companies, such as CytoReason, are specifically focused on building advanced disease models, for example.

In March 2023, AstraZeneca presented preclinical data on an AI-generated target, the Serum Response Factor (SRF), for idiopathic pulmonary fibrosis (IPF) — from its collaboration with UK-based AI company BenevolentAI. The target, discovered via BenevolentAI’s AI-enabled drug discovery engine, underwent thorough experimental validation by AstraZeneca, involving CRISPR screening in primary human lung fibroblasts and validation via SRF gene silencing or pharmacological SRF pathway inhibition. The presented data indicates that inhibiting SRF-driven transcription of pro-fibrotic genes in lung fibroblasts could potentially lead to antifibrotic efficacy in IPF. To date, the collaboration between BenevolentAI and AstraZeneca has resulted in five AI-generated targets selected for portfolio entry, three of which are for IPF. This successful partnership was expanded in January 2022 for another three years, including two new disease areas – systemic lupus erythematosus and heart failure.

Several months earlier, AstraZeneca announced a strategic research collaboration with Illumina, a global pioneer in DNA sequencing and array-based technologies. This collaboration aims to expedite drug target discovery by melding their respective competencies in AI-based genome interpretation and genomic analysis. The initiative will examine if a unified approach utilizing these technologies can bolster the efficiency and certainty of target discovery in pursuit of promising drugs built upon human omics insights. AstraZeneca’s Centre for Genomics Research will adopt a framework merging the AI-based tools of both companies, leveraging next-generation AI interpretation tools like Illumina’s PrimateAI and SpliceAI, along with AstraZeneca’s own tools such as JARVIS and in silico predictors.

In September 2022, Pfizer announced the expansion of its multi-year partnership with Israel-based AI in pharma company CytoReason. Under this agreement, Pfizer will invest $20M in equity, with the option to license CytoReason’s platform and disease models and fund further project support in a deal that could reach up to $110M over the next five years. Since the initiation of the collaboration in 2019, Pfizer has utilized CytoReason’s biological models in its research to boost the understanding of the immune system for the development of drugs for immune-mediated and immuno-oncology diseases. This additional investment will aid the development of more disease models and the creation of high-resolution models across various therapeutic areas.

CytoReason is also a partner for another big pharma, Sanofi. In January 2023, the two companies announced expansion of collaboration, utilizing CytoReason’s AI platform for inflammatory bowel disease (IBD) target discovery. This deal will support Sanofi’s efforts to identify IBD patient subtypes and match them with relevant targets. This partnership follows a project initiated in 2021, where CytoReason’s cell-centered models were used to provide insights for asthma endotypes. The extended agreement involves a substantial but undisclosed financial commitment from Sanofi. 

Last year, Sanofi also announced a multi-year, multi-target research collaboration with Hong Kong-based Insilico Medicine, leveraging the latter’s Pharma.AI platform to expedite drug discovery. Insilico, a pioneer in applying deep learning for drug discovery, will assist Sanofi in developing treatments in areas such as cancer, fibrosis, and immunity. The collaboration signifies a $21.5m investment by Sanofi for upfront and target nomination fees, granting access to Insilico’s AI platform and their interdisciplinary team of scientists. The partnership holds potential for further payments up to $1bn if key R&D and sales milestones are reached.

AI for target based and phenotypic drug discovery

Another popular and promising use case of applying AI in the pharmaceutical industry is for drug design and lead optimization. 

In one illustrative example, Sanofi initiated a strategic research collaboration with San Francisco-based Atomwise, a pioneer in applying artificial intelligence for screening small molecules. This alliance is set to exploit Atomwise’s AtomNet® platform, employing its computational discovery capabilities to investigate up to five drug targets provided by Sanofi. Atomwise’s platform integrates deep learning in structure-based drug design, providing an AI-fueled search in its proprietary library of over 3 trillion synthesizable compounds. Sanofi’s investment in this partnership includes an upfront payment of $20 million to Atomwise for the identification, synthesis, and further development of lead compounds, with potential additional payments surpassing $1 billion tied to critical research, development, and sales milestones, as well as tiered royalties. This collaboration is expected to catalyze the discovery of new treatments for diseases that have previously been challenging due to elusive or inadequately characterized drug targets.

US-based big pharma AbbVie has entered into an antibody discovery agreement with Canadian AbCellera, reinforcing AbCellera’s recent string of partnerships. The collaboration aims to develop antibody candidates for up to five targets across various indications. AbbVie plans to harness AbCellera’s AI-driven antibody discovery and development engine, taking responsibility for the development and commercialization of any antibodies discovered during their collaboration. The agreement stipulates that AbCellera is entitled to research payments, as well as clinical, commercial milestone payments, and royalties. While specific details regarding the targeted timeline or indications were sparse, AbCellera’s CEO, Carl Hansen, commented to BioSpace that their discovery and development engine is designed to overcome the limitations of conventional discovery methods, aiming to identify optimal clinical candidates with increased precision and speed

In March 2023, Eli Lilly announced collaboration with XtalPi, an AI in pharma company, on a $250m project. The collaboration will exploit XtalPi’s AI and robotics platform for the de novo design and delivery of drug candidates for an undisclosed target. XtalPi’s integrated capabilities in AI and robotics will be utilized to create a novel compound, which will then be advanced by Eli Lilly through clinical and commercial development. XtalPi’s ID4Inno platform, designed for small-molecule drug discovery, will be key in creating a target-specific mega chemical space and identifying promising lead series. The synthesized molecules will be tested using XtalPi’s internal biochemical, pharmacodynamic, cellular, and pharmacokinetic assay capabilities. These capabilities, paired with XtalPi’s multiple autonomous robotic workstations, highlight the value of leveraging AI in pharmaceuticals for energy-efficient, precise parallel chemical synthesis and assays. 

Exscientia has seen their partnership with Bristol Myers Squibb (BMS) yield significant returns, advancing the first of three candidates designed for first-in-human trials this year. The collaboration, which could generate over $1.3 billion for Exscientia, has resulted in the production of EXS4318, a novel immunology and inflammation (I&I) small molecule that will be overseen by BMS through Phase I trials.

EXS4318, a potential first-in-class selective Protein kinase C (PKC) theta inhibitor, sprouted from the initial AI-based small molecule discovery collaboration launched by Exscientia and Celgene back in March 2019. Following BMS’ $74 billion acquisition of Celgene later that year, the collaboration was expanded to include I&I and oncology candidates, significantly increasing the potential earnings for Exscientia.

BMS’ commitment to the partnership is underscored by their agreement to potentially pay more than $1.3 billion in clinical, regulatory, and commercial payments, which include up to $50 million upfront and up to $125 million in “near to mid-term” milestones, as well as tiered royalties on net sales.

In December 2021, Recursion Pharmaceuticals, a Utah-based clinical-stage AI-driven ‘digital biotech’ company, became a target of interest for Roche and Genentech (a member of the Roche Group). The companies announced collaboration in neuroscience and oncology, aiming to advance medicines using machine learning and target-agnostic high-content screening methods. The collaboration, worth several billion dollars, exploits Recursion’s technology-enabled drug discovery platform, the Recursion Operating System (OS). Under the agreement, Recursion will receive an upfront payment of $150 million and may earn additional performance-based research milestones. The Recursion OS combines wet-lab and dry-lab biology, enabling the industrialization and digitization of drug discovery. 

Since 2020, another prominent big pharma, Pfizer, has been leveraging IBM’s supercomputing and AI capabilities to facilitate the creation of new medications, such as PAXLOVID, an oral treatment for COVID-19 that received approval in 2022. Pfizer maintains that this technology has trimmed computational time by 80-90%, asserting that it expedited the drug design process to a mere four months.

AI for designing better clinical trials 

Clinical trials are a notorious bottleneck of the entire drug development route. It is in clinical trials that many promising drug candidates with excellent preclinical data fail, rendering huge costs and lost opportunities for patients and for big pharma. Smart and data-driven clinical trial design is essential for increased success of a drug candidate getting to FDA approval. Artificial Intelligence is playing an increasingly important role in designing clinical trials, including biomarker discovery, prediction of treatment responses, and optimization of trial protocols, allowing for more precise patient selection and reducing the overall cost and duration of trials. AI also enables real-time monitoring and adaptive trial designs, enhancing the flexibility and responsiveness of clinical studies.

For instance, according to a 2021 study ‘Does biomarker use in oncology improve clinical trial failure risk? A large‐scale analysis’, clinical trials designed without integration of relevant and informative biomarkers are 12 times more likely to fail.

To illustrate this trend, let’s start with UK big pharma GlaxoSmithKline (GSK), which is actively engaging in AI-driven partnerships to boost its clinical trial design capabilities. In March 2023, GSK announced its collaboration with AI in pharma company PathAI on a randomized Phase 2b clinical trial called HORIZON, which focuses on non-alcoholic steatohepatitis (NASH). The trial aims to evaluate improvements in liver histology using GSK4532990 compared to a placebo in participants with NASH and advanced fibrosis. PathAI’s role in the collaboration involves generating, digitizing, and analyzing liver biopsy slides for evaluation by pathologists. The company will also utilize its AI-based Measurement of NASH Histology (AIM-NASH) tool to provide histologic evaluation and generate exploratory endpoints for the study. PathAI’s end-to-end anatomical pathology services, including kitting, logistics, and lab and analytical services, will be utilized through its Biopharma Lab in Memphis, TN. The AIM-NASH tool has been trained to detect and quantify key histological features of NASH, offering a comprehensive assessment of disease severity. The collaboration builds on PathAI and GSK’s existing partnership in NASH and oncology research and drug development.

Last year, GSK entered into a three-year partnership with Tempus, an AI-driven tech vendor, to enhance clinical trial design, subject enrolment, and drug target identification. Through the collaboration, GSK will gain access to Tempus’ AI-enabled platform, which includes a library of de-identified patient data. By leveraging Tempus’ platform and combining it with GSK’s expertise in Artificial Intelligence and Machine Learning (AI/ML), the partnership aims to improve GSK’s R&D success rate and enable faster, tailored treatments for patients. The partnership involves a three-year financial commitment, with an initial payment of $70 million from GSK, and GSK holds an option to extend the deal for another two years. This collaboration builds upon the existing partnership between GSK and Tempus, which started in 2020 and focused on clinical trial enrolment for specific cancer types. GSK expects the collaboration with Tempus to provide unique insights for discovering better medicines and transforming drug discovery processes. Currently, the companies are collaborating on an open-label Phase II trial, utilizing a data-guided approach to expedite study timelines and optimize site selection and subject enrolment. This collaboration follows positive findings from GSK’s Phase II PERLA trial, evaluating Jemperli (dostarlimab) plus chemotherapy in patients with first-line, metastatic, non-squamous non-small cell lung cancer.

In January 2023, Tempus announced a prospective study, in collaboration with AstraZeneca, to identify biomarkers of response in patients with small cell lung cancer (SCLC). The study, titled Sculptor, is co-sponsored by Tempus and AstraZeneca’s Personalize SCLC Initiative and is currently open for enrollment. SCLC is an aggressive disease with limited therapeutic targets, and there is a high unmet need for effective treatments. The Sculptor study aims to leverage Tempus’ molecular profiling offerings to gather insights that can support early research and identify distinct patient segments that may benefit from emerging therapies. The study is currently active at five TIME Trial Network sites and plans to expand to additional providers across the United States. This collaboration represents a precision medicine approach to improve overall survival rates and advance the understanding of SCLC.

In June 2022, Bristol Myers Squibb (BMS) partnered with France and US-based AI developer Owkin to leverage artificial intelligence in the design and optimization of cardiovascular drug trials. The partnership aims to enhance clinical trial design and execution by using machine learning techniques, such as optimized endpoint definitions, patient subgroup identification, and treatment effect estimation. Venkat Sethuraman, SVP of Global Biometrics and Data Sciences at BMS, highlighted the company’s use of AI in clinical trials, including the simulation of trials using existing datasets, digitally enabled trials using real-world data, and automation of processes through robotics. The collaboration with Owkin will specifically focus on optimizing clinical trials in cardiovascular disease, potentially eliminating placebo arms in some trials and improving efficiency in trials for rare diseases.

In September 2021, AstraZeneca established a strategic partnership with Oncoshot to enhance the recruitment process for cancer clinical trials in Singapore. Oncoshot’s digital platform, InSite Feasibility, utilizes data analytics and real-time insights to streamline oncology feasibility studies. Powered by Artificial Intelligence (AI)-enabled patient-to-trial matching technology, the platform translates de-identified data from Singapore’s cancer population into precise analytics, facilitating cancer research. The partnership allows AstraZeneca to swiftly initiate trials that are most relevant to Singapore’s cancer population. With Singapore being a leading cancer research hub, this collaboration addresses the inefficiencies in current feasibility studies and aims to optimize patient enrollment. By leveraging accurate data analytics, AstraZeneca and Oncoshot seek to accelerate clinical trials, provide innovative medicines, and contribute to Singapore’s innovation initiatives in healthcare.

AI for drug repurposing programs 

Drug repurposing is one of the golden mines for AI-based technologies to drive value since a lot of data is already known about the drug in question. Repurposing previously known drugs or late-stage drug candidates towards new therapeutic areas is also a desired strategy for many biopharmaceutical companies as it presents less risk of unexpected toxicity or side effects in human trials, and, likely, less R&D spend.    

Leveraging insights derived from BenevolentAI’s AI system, Eli Lilly repurposed its rheumatoid arthritis drug, baricitinib, for an alternative indication as a potential treatment for COVID-19. In early 2020, BenevolentAI utilized its advanced AI system to propose baricitinib, owned and marketed by Eli Lilly under the brand name Olumiant™, as a potential therapeutic option against the virus. Following this AI-derived hypothesis, the FDA endorsed an Emergency Use Authorization for baricitinib in the treatment of hospitalized COVID-19 patients in November 2020, following promising phase III data in hospitalized Covid-19 patients. The ACTT-2 trial data further validated this proposition, indicating an improvement in the clinical outcomes and a 35% reduction in mortality rate among patients treated with baricitinib, although the drug failed to meet the trial’s primary endpoint — reducing the risk of the disease worsening. This swift transition from the AI-proposed hypothesis to clinical trials and subsequent FDA.

AI for developing drug formulations

AI in the pharmaceutical industry appears to also be useful for drug development, including optimization of drug formulations. 

In April 2023, Merck and XtalPi Inc. announced collaboration on a study showcasing the benefits of combining computational workflows with wet lab experiments in drug development. The study focused on the impact of polymer additives on the crystal habit of metformin HCl, a diabetes medication. By utilizing XtalPi’s morphology prediction platform and Merck’s experimental capabilities, a comprehensive screening approach for crystal morphology engineering was developed. XtalPi’s molecular dynamics predictions successfully predicted the influence of polymer additives on metformin HCl’s crystal habits, with experimental observations confirming the transformation of crystal morphology. This collaboration represents a ‘digital-first’ approach that combines computer simulations and experimental formulation expertise to enhance pharmaceutical development processes.

Connecting the dots with AI

One of the strongest aspects of artificial intelligence (deep learning, NLP) is the ability to integrate multimodal data from various sources, presenting insights about the system in general, taking into account multiple processes and networks — both scientific, and operational. 

In this context, Sanofi is making strides in incorporating artificial intelligence (AI) into its research and development activities, starting with the rollout of its AI app, Plai. Developed in partnership with Aily Labs, Plai provides real-time data and a comprehensive view of all Sanofi activities, offering personalized insights and scenarios to support staff. This is part of Sanofi’s broader ambition to become the first pharmaceutical company powered by AI at scale, equipping its workforce with tools and technologies that enhance decision-making. AI is already benefiting Sanofi’s drug discovery, clinical trial design, and manufacturing processes, and the company believes there is still untapped potential for AI in healthcare. For instance, AI has accelerated research processes from weeks to hours and improved target identification in therapeutic areas like immunology, oncology, and neurology. Sanofi’s collaborations with AI startups, including Insilico Medicine, Exscientia, and Owkin, further demonstrate its commitment to leveraging AI for drug discovery. The company also plans to utilize Plai in its clinical trial operations to enhance enrollment, particularly among underrepresented populations.

Conclusions: R&D outsourcing and M&A activity will be increasing in “AI for drug discovery” space

With an increasing interest in AI-driven technologies among the leading pharmaceutical and biotech companies, a strategic focus of pharma and biotech businesses will be further shifting towards AI partnerships, R&D outsourcing, and M&A activity as means to quickly get access to the required expertise and know-hows. Complex nature of AI-based technologies, a need for costly and sophisticated IT infrastructure, a fast pace of progress in the field, and a relative scarcity of highly skilled data science specialists to support specialized machine learning research — these are some of the key drivers of the ascending outsourcing trend.

Topics: Emerging Technologies    Industry Trends   

• SEO Powered Content & PR Distribution. Get Amplified Today.
• PlatoData.Network Vertical Generative Ai. Empower Yourself. Access Here.
• PlatoAiStream. Web3 Intelligence. Knowledge Amplified. Access Here.
• PlatoESG. Automotive / EVs, Carbon, CleanTech, Energy, Environment, Solar, Waste Management. Access Here.
• BlockOffsets. Modernizing Environmental Offset Ownership. Access Here.
• Source: https://www.biopharmatrend.com/post/34-biopharmas-hunt-for-artificial-intelligence-who-does-what/