AI Drug Discovery Summer ๐๏ธ News
"Things don't turn up in this world until somebody turns them up". By James A. Garfield
Welcome back ๐ค to another edition of MetaphysicalCells.
AI in Drug Discovery market in terms of revenue was estimated to be worth $0.6 billion in 2022 and is poised to reach $4.0 billion ๐ฐ by 2027, growing at a CAGR of 45.7% from 2022 to 2027 according to a latest report published by Markets and Markets.
The growing need to curb the drug discovery cost & reduce the overall time taken in this process, the rising adoption of cloud-based applications and services, and the impending patent expiry of blockbuster drugs are some of the key factors driving the growth of this market. You can download ๐ฉ the report here.
Major Milestones in AI Drug Discovery
๐ In 2020, Exscientia announced that the first-ever ๐ฅ AI-designed drug molecule (DSP-1181) in collaboration with Sumitomo Dainippon Pharma, had entered human clinical trials. DSP-1181 was developed in only 12 months (instead of 5 years the average โex normalโ), in order to treat obsessive-compulsive disorder.
Right now, three years later, more than ten AI-designed drugs ๐ are already in or are entering clinical trials:
REC-2282 (Neurofibromatosis type 2), REC-994 (Cerebral cavernous malformation), REC-4881 (Familial adenomatous polyposis), Undisclosed (HRD-negative ovarian cancer) by Recursion Pharmaceuticals (Founded 2013, Salt Lake City, Utah),
INS018_055 (Idiopathic pulmonary fibrosis) by InSilico Medicine (Founded 2014, Hong Kong, China),
BEN-2293 (Atopic Dermatitis), BEN-8744 (Ulcerative colitis) by BenevolentAI (Founded 2013, London, UK),
EXS-21546 (Solid tumours carrying high adenosine signatures) by Exscientia (Founded 2012, Oxford, UK) and Evotec,
LP-184 (Pancreatic cancer, malignant gliomas and atypical teratoid rhabdoid tumours) by Lantern Pharma (Founded 2013, Texas US),
VRG50635 (Amyotrophic lateral sclerosis) by Verge Genomics (Founded 2015, California, US),
OPL-0401 (diabetic retinopathy) by Valo Health (Founded: 2019, Boston, Massachusetts)
RLY-4008 (FGFR2-altered cholangiocarcinoma) by Relay Therapeutics (Founded 2015, Massachusetts, US) and
EXS-4318 (Inflammatory and autoimmune conditions) by Exscientia in collaboration with Bristol Myers Squibb.
๐ In 2021, the AI system AlphaFold by DeepMind ๐๐ฅ predicted the protein structures for 330,000 proteins, including all 20,000 proteins in the human genome. In this article โOne of the Biggest Problems in Biology Has Finally Been Solved", Google DeepMindโs CEO Demis Hassabis explains how its AlphaFold AI program predicted the 3-D structure of every known protein.
๐ In June 2022, Exscientia announced preliminary results ๐ from a phase 1 trial of the small molecule EXS-21546 (a highly selective adenosine A2A receptor antagonist) co-developed with Evotec. EXS-21546 has subsequently entered phase 1b/2 trials for patients with solid tumours carrying high adenosine signatures. Exscientia has raised a total of $673.7M over 12 rounds.
๐ At the end of 2022, BenevolentAI completed a phase 2a trial for BEN2293, a topical ointment for the treatment of atopic dermatitis but even if the treatment was found to be safe did not meet โ its secondary endpoint of reducing itch and inflammation, according to a company press release in April 2023.
In 2018, BenevolentAI technically reached โunicornโ ๐ฆ status when it achieved a valuation of just over $1.7B after a big fundraising. However, the companyโs valuation suffered significantly when Neil Woodford, their major investor, was forced to wind up his fund in 2019. In December last year, BenevolentAI merged with Odyssey Acquisition, an Amsterdam-listed special-purpose acquisition company, in order to list on the Euronext Amsterdam stock exchange. Since the merger, BenevolentAIโs valuation is almost back to 2018 levels at $1.5B
๐ In January 2023, AbSci became the first company ๐ฅ๐ โto create and validate de novo antibodiesย in silicoโ using generative AI. In April 2023, AbSci and M2GEN (now Aster Insights), an oncology-focused health informatics solutions company with the most advanced lifetime consented clinico-genomics data, announced a partnership to create new cancer medicines. In particular, Absciโs generative AI drug creation platform will tap into M2GENโs clinical and molecular data set, ORIEN AVATAR, to accelerate the creation of therapeutics. Absci has raised a total of $237.9M in funding over 10 rounds.
๐ In January 2023, Insilico Medicine announced an encouraging topline readout of its phase 1 safety and pharmacokinetics trial of INS018_055, for idiopathic pulmonary fibrosis (IPF). FDA has granted INS018_055 orphan drug designation for the treatment ofย IPF and Insilico is expecting to initiate a global phase 2 clinical trial testing INS018_055 in people with IPF later this year. INS018_055, a protein kinase inhibitor like ISM042-2-048, is the first drug discovered and designed using AI ๐ฅ๐๐๏ธto reach the phase I clinical trial milestone. Insilico Medicine has raised a total of $401.3M in funding over 10 rounds.
๐ In June 2023, Verge Genomics announced the completion of the Phase 1 clinical trial of VRG50635 in healthy adult volunteers. Detailed results ๐๏ธ๐ from the study will be presented at the upcoming European Network to Cure ALS (ENCALS) 2023 meeting being held in Barcelona, Spain from July 12-14, 2023. Verge Genomics has raised a total of $134.1M in funding over 4 rounds.
๐ In June 2023, Lantern Pharma (an AI drug repurposing company) received FDA clearance of IND Application for the drug candidate LP-184 in solid tumours, and the first-in-human phase 1A clinical trial is anticipated to launch this summer for multiple advanced solid tumours and brain and central nervous system cancers. Lantern has two more clinical candidates LP-300 and LP-100 and has raised a total of $95M.
Source:
Inside the nascent industry of AI-designed drugs By Carrie Arnold,ย
How Artificial Intelligence is Revolutionising Drug Discovery By Matthew Chun, and
These Six Biotechs are Winning the Race to Get AI-Designed Drugs to the Clinic By Helen Albert
AI Drug Discovery News ๐ฐ
๐ฐ Sanofi has just declared its ambition to incorporate AI and data science across all its operations. In particular, through its innovative AI app plai, Sanofi will be able to make real-time, reactive data interactions possible and have a comprehensive view of the companyโs activities from research to clinical operations to manufacturing and supply to business analysis (Sanofi Goes โAll Inโ to Lead the AI Revolution in Pharmaceutical Industry).
The app plai was developed in collaboration with the AI platform company Aily Labs, and delivers real-time, reactive data interactions while can give an unprecedented 360ยฐ view across all Sanofi's activities. By aggregating available company internal data across all functions and by utilising AI the app provides timely insights and personalised โwhat ifโ scenarios to support thousands of Sanofi teams decision makers to take informed decisions in a simple and modern digital user experience.
Moreover, in 2022 Sanofi acquired Amunix Pharmaceuticals which uses AI to tailor-deliver medicines that become active only in tumour tissues, while not harming normal ones. The same year Sanofi joined forces in a collaboration ๐คwith Exscientia (committing potentially up to $5.2 billion to develop up to 15 novel small molecule candidates across oncology and immunology), with Insilico Medicine (signing a strategic research worth up to $1.2 Billion) and with Atomwise (inking a $1.2 billion biobucks research collaboration). This comes on top of Sanofiโs partnership with Owkin in 2021 (in a $270M cancer AI deal), to utilise Owkinโs AI-driven platform.
๐ฐ A group of researchers at the Ohio State University has created an AI framework called G2Retro to automatically generate reactions for any given molecule. This new AI framework was able to cover an enormous range of possible chemical reactions as well as accurately and quickly discern which reactions might work best to create a given drug molecule (G2Retro as a two-step graph generative models for retrosynthesis prediction). G2Retro is one-step retrosynthesis prediction tool that imitates the reversed logic of synthetic reactions.
๐ข โOur generative AI method G2Retro is able to supply multiple different synthesis routes and options, as well as a way to rank different options for each molecule. This is not going to replace current lab-based experiments, but it will offer more and better drug options so experiments can be prioritised and focused much faster.โ
By Xia Ning, PhD, associate professor of computer science and engineering at Ohio State University
๐ฐ On June 15, 2013, it was announced that algorithms have led to the discovery of three drugs that have the potential to delay the effects of aging ๐ด๐ต๐ง (AI and anti-aging research: Unveiling the latest drug discovery). In particular, a group of researches from The University of Edinburgh have developed an innovative method that employs AI to identify senolytic drugs, by leveraging data from over 2,500 chemical structures extracted from past studies, and training a ML model to recognise the essential characteristics associated with chemicals possessing senolytic activity.
๐ฐ So far, the large language model ChatGPT that belongs to a family of AI known as generative AI has been used during drug discovery ๐ข
for data analysis, literature reviews, virtual screening, predictive modelling, and decision support. Chat GPT-4 in particular seems to be very useful in aiding the drug discovery process,
was customised by Insilico Medicine as โChatPandaGPTโ and integrated into its PandaOmics platform, which enables researchers to have natural language conversations with its platform and efficiently navigate and analyse large data sets, facilitating the discovery of potential therapeutic targets and biomarkers, andย
can be integrated (like Synthace did) into platforms of other companies in order to design protocols for biology experiments and automate lab work (Using ChatGPT to design and automate biology experiments).ย
Source: Can ChatGPT be used to advance drug discovery?
๐ฐ XtalPi just made a partnership ๐พ deal with US-based Eli Lilly, to leverage AI for drug discovery. The two companies will utilise XtalPiโs integrated AI capabilities and robotics platform for the de novo design and delivery of drug candidates for an undisclosed target (XtalPi announces $250m AI drug discovery collaboration with Eli Lilly).
โWith a closed loop of AI and quantum physics algorithms working in sync with the data factory of large-scale robotics experiments, XtalPi is uniquely equipped to tackle challenging novel targets. We are honoured that Lilly has chosen XtalPiโs AI plus robotics drug research and development platform as a partner in achieving more fruitful pharmaceutical innovation and bringing much-needed treatments to patients worldwide.โ
By XtalPi's CEO Dr Jian Ma
Shenzhen-based Xtalpi is backed by Sequoia and has raised a total of $786.4M.
๐ฐ The University of Sydney ๐ฆ just signed a memorandum of understanding (MoU) with Pharos Therapeutics, the Australian subsidiary of South Korean pharmaceutical company Pharos iBio (AI drug discovery will power collaboration with Pharos Therapeutics). The MoU will provide access for the Universityโs Drug Discovery Initiative to Pharos iBioโs proprietary AI drug development platform, Chemiverse, while Pharos will benefit from working with the Universityโs world-class team of researchers.
๐ฐ Researchers at MIT and Tufts University have devised an alternative computational approach based on a large language model for drug screening ๐experiments. This large language model, that can analyse huge amounts of text and figure out which amino acids are most likely to appear together, is known as ConPLex and can match target proteins with potential drug molecules without having to perform the computationally intensive step of calculating the moleculesโ structures (โContrastive learning in protein language space predicts interactions between drugs and protein targetsโ). In other words, the approach co-locates the proteins and potential drug molecules in a shared feature space while learning to contrast true drugs from similar nonbinding โdecoyโ molecules. ConPLex is extremely fast, which allows it to rapidly shortlist candidates for deeper investigation.
๐ฐ IndyGeneUS, a black-owned ๐ฉ and service disabled veteran-owned and operated small business (plus a Merck Digital Sciences Studio portfolio company) that provides a decentralised, blockchain-secured platform for patients to confidentially store and analyse their genetic data, has just teamed up with AI Squared to accelerate ML adoption across its organisation and improving decision-making with AI-driven innovation (Tech Start-ups IndyGeneUS and AI Squared Forge Partnership to Drive Health Innovation Through Advanced AI Technologies).
Until next time ๐ชผ๐ฆ๐ชธ,