AI-powered drug discovery: update (I)
AI-powered drug discovery: update (I)
MetaphysicalCells: A newsletter about Science, Technology and AI Drug Discovery
This article is part of a series of articles dedicated to AI-drug discovery progress of drug candidates in preclinical and clinical phase.
“In nature nothing is created, nothing is lost, everything changes."
By Antoine Lavoisier (Traité élémentaire de chimie)
Index for “AI-powered drug discovery: update part I” 📇
Deep Genomics
Schrödinger
Biotx.ai
NeoCura
Healx
Octant, Inc
Delta4 AI
Model Medicines
Modulus Co
Atomwise
DeepCure
Hummingbird Biosciences
Deep Genomics Inc
Deep Genomics (2015, Canada) has a proprietary AI platform, BigRNA, that is the world’s first RNA foundation model for RNA therapeutics that can predict tissue-specific RNA expression, splicing, microRNA sites and RNA binding protein specificity. BigRNA can uniquely discover a wide range of new biological mechanisms and RNA therapeutic candidates that would not be found using traditional approaches that measure only overall gene expression levels. In contrast, BigRNA is trained to predict RNA expression at sub-gene resolution, such as polyadenylation, exon skipping and intron retention (An RNA foundation model enables discovery of disease mechanisms and candidate therapeutics).
Deep Genomics made headlines back in 2019 with discovering a novel target and a novel RNA therapeutics candidate, DG12P1, for rare Wilson disease using their AI workbench platform BigRNA, all within 18 months of initiating target discovery efforts (Deep Genomics Nominates Industry’s First AI-Discovered Therapeutic Candidate). In particular,
Deep Genomics’ AI system scanned over 2,400 diseases and over 100,000 pathogenic mutations while searching for good drug development opportunities and was able to predict and confirm the precise disease-causing mechanism of the mutation Met645Arg. One of several genetic mutations that leads to loss of function of the ATP7B copper-binding protein, a genetic mutation that impairs the body’s ability to remove copper, and thereby identify a clear therapeutic target. The AI system was then used to identify 12 lead candidates out of thousands of potential compounds, taking into account in vitro efficacy and toxicity. In particular, TDG12P1 was designed to correct the exon skipping mechanism of Met645Arg and after tolerability experiments Deep Genomics declared it the ideal candidate to advance toward IND.
🎳 On June 12, 2024, Deep Genomics announced the opening of its new office and lab facility in Cambridge, Massachusetts (the expansion of its Toronto office) and several key leadership hires. Deep Genomics raised so far a total of $180M to automate drug discovery.
Schrödinger Inc
Schrödinger (1990, New York US) is involved in computational drug design for pharmaceutical and biotechnology research, and provides platforms for drug discovery, materials science and molecular modeling, with a physics based chemical simulation software that accelerates the convergence of physics, ML and enterprise informatics. Built upon more than 30 years of R&D, their industry-leading computational platform is transforming the way therapeutics and materials are discovered by enabling highly accurate in silico predictions of key molecular properties across vast chemical space.
Back in 2022, Schrödinger secured an FDA green light to study its computer-designed therapy for non-Hodgkin lymphomas in an early-phase trial, SGR-1505, a potential drug to block MALT1 (the Mucosa-associated lymphoid tissue lymphoma translocation protein 1 or paracaspase), a crucial protein for T and B lymphocytes activation.
In order to select SGR-1505, the company’s ML platform sorted through 8.2 billion potential compounds over a period of 10 months, identifying initially 78 compounds that went through preclinical experiments and finally they selected the most promising candidate: SGR-1505.
In 2023, Schrödinger announced new preclinical data regarding SGR-1505 (and SGR-2921, its investigational CDC7 inhibitor) that showed that both SGR-1505 and SGR-2921 have multiple favorable attributes and the potential for combination activity with standard-of-care agents.
Schrödinger also reported preliminary pharmacodynamic data from the Phase 1 study of SGR-1505 in healthy subjects, which showed that SGR-1505 inhibited cytokine release in human whole blood and demonstrated evidence of MALT1 inhibition in humans.
Apart from its proprietary pipeline, Schrödinger’s therapeutics group is working on a number of collaborative drug discovery programs with Ajax Therapeutics, Bright Angel Therapeutics, Bristol Myers Squibb, Lilly, Loxo Therapeutics, Morphic Therapeutic, Nimbus Therapeutics, Otsuka and Sanofi. For example:
In 2020, Schrödinger announced an expanded collaboration with AstraZeneca focused on refining a biologics modeling solution with the aim of speeding up the development of antibody and protein-based therapeutic candidates.
In collaboration with Loxo Therapeutics, Schrödinger is progressing an undisclosed program in oncology currently in Phase I clinical studies. And
On January 08, 2020, Schrödinger and Bayer Collaborated to Co-Develop de novo Design Technology to Accelerate Drug Discovery for enumerating, screening and scoring billions of synthetically feasible, virtual compounds for the identification and optimization of potential new therapeutic candidates. The de novo design technology will integrate the power of Schrödinger’s molecular design technology with Bayer’s proprietary in silico models predicting compound ADMET and chemical synthesizability. The result of this collaboration is a drug design platform that they termed at Bayer the NovaDesign platform.
🏓 On May 1, 2024, the company revealed a challenging first quarter with total revenue falling significantly year-over-year contrary to analyst expectations (Schrodinger Inc Misses Revenue Expectations in Q1 2024). Moreover, according to Nanalyze, Schrödinger (NASDAQ: SDGR) may now see negative growth based on the middle of their 2024 guidance (Is Schrödinger Stock Coming Under Pressure From AI?). However, Schrödinger is still a Top 5 AI & Digital Biotech Companies 🎩 (June 2024).
Schrödinger has raised a total of $567.2M.
Biotx.ai GmbH
Biotx.ai (2017, Berlin Germany) delivers predictive biomarkers as a basis for decision support in precision medicine. They use novel, state of the art, ML approaches specifically designed to find complex patterns in biomedical data. Traditionally, big data analysis has involved millions of subjects and few features that are easy to train AI with. But, with very small patient cohorts and sample sizes (amount of information gathered through clinical trials and genetic testing), with few subjects (few patients) and millions of features (an incredibly large number of data points, like the entire genome) is hard to train AI with. This is something Biotx.ai calls “wide data”, that is incredibly hard to analyze consequently most pharmas avoid this approach entirely, opting instead to sequence entire populations. For this reason, biotx.ai has designed its AI algorithms to make wide data manageable by separating meaningful findings from the noise and leading to the discovery of previously undetectable complex genetic patterns (What is Wide Data?).
They offer Causal inference modeling at scale that requires the researcher to construct a model to explain the relationships among concepts related to a specific phenomenon, and is a proven method to ensure success in clinical trials. They enable drug discovery, drug repurposing and validation via large-scale screening:
Discovering all diseases that can be treated by a given drug,
Discovering all drugs that can be used to treat any given disease and
Exploring all disease mechanisms and pathways.
By utilizing Biotx.ai
Immungentics, a German biotech focused on Alzheimerʼs disease, was able to find causal biomarkers crucial for their study design and used multiple causal models from biotx.aiʼs to better understand the mechanism of action of their drug thiethylperazine via the ABCC1 transporter.
Alfa Intes, a family-owned Italian pharma, discovered new indications for its drug Indomethacin. And
Eternygen, a German biotech, found that one of the potential indications for their compound will not work and has shifted focus onto another indication instead by signing a next deal.
On October 26, 2023, Simbec-Orion, a full-service mid-size CRO, and Biotx.ai announced their strategic partnership to bring AI-powered insights and predictive models to support data-driven clinical trial design.
On August 21, 2023, Iceland-based 3Z announced a strategic partnership with Biotx.ai to improve the transition from animal studies to human clinical trials. The study will utilize AI-powered human genetics-based modeling and explore the involvement of 3Z’s lead therapeutic targets in ADHD.
Biotx.ai raised a total of $40K.
NeoCura/ 新合概况 or Shenzhen Xinhe Bio-Medical Technology Co Ltd
NeoCura (2017, China) is a high-tech enterprise that relies on AI technology to create an RNA technology platform and conduct R&D of innovative drugs. The company has built multi-omics big data acquisition platforms and multiple omics databases, to conduct in-depth drug target mining and fully automated drug design. So far, NeoCura has built 1) a unique Drug Screening & Validation platform for neoantigens. 2) a Target Mining platform and 3) an RNA Drug platform for delivery (LNP delivery system and new delivery system), multi-omics (NGS, ribo-Seq, SHAPE-seq) and for nucleotide modifications.
They have several candidates (cancer vaccines, virus vaccines and tumor immunomodulator) in preclinical phase, and as of 2024 two candidates in clinical development:
Personalized Neoantigen Tumor Vaccine, PGV-002, is currently in Phase II for Solid Tumor.
XH-001 is under clinical development for the treatment of advanced solid tumors. It is an mRNA-based tumor neoantigen vaccine with a likelihood of approval.
In 2023, NeoCura and Beijing GenePlus Technology Co., Ltd ("Gene+") signed a strategic cooperation agreement that will leverage Gene+'s testing platform 🛠️ technology to accelerate the early clinical trial progress of mRNA tumor neoantigen vaccines.
NeoCura has established leading R&D centers and production centers in Beijing, Shenzhen, and Guangzhou to support pipeline development, platform upgrades and clinical demand. And has raised a total of $116.97M.