Hi! I am Andrii Buvailo, and this is my weekly newsletter, ‘Where Tech Meets Bio,’ where I talk about technologies, breakthroughs, and great companies moving the biopharma industry forward.
If you've received it, then you either subscribed or someone forwarded it to you. If the latter is the case, subscribe by pressing this button:
In today’s newsletter:
Weekly tech+bio highlights;
Seed Health Invests in AI-Driven Biology Platform to Enhance Microbiome Science;
Novel Gen AI Platform for Preclinical Research;
This Company's ML Model Predicts KRAS Inhibitor's Clinical Benefit in Patients with >80% Accuracy (paid);
Model Medicines Advances AI Drug Discovery in Antiviral and Oncology Fields (paid);
A Novel Quantum-assisted Approach to Designing Drugs From Fragments (paid);
Cancer Vaccines: insights from #AACR24
Mapping AI-driven drug candidates and targets (infographic)
This week’s sponsor: Trialize
In the realm of clinical trials, efficient data management and monitoring are crucial yet often hampered by outdated and time-consuming practices.
Trialize's innovative approach to automating study builds and leveraging artificial intelligence (AI) for data analysis significantly reduced the time from protocol to trial, transforming the process from a traditionally months-long undertaking into a matter of days.
Trialize’s architecture offers a comprehensive, integrated solution covering all aspects of clinical trial management, from Electronic Data Capture (EDC) and Electronic Patient-Reported Outcomes (ePRO) to TeleVisits, eConsents, and Analytics. Its interface is designed to be intuitive and user-friendly, drastically reducing the learning curve for study teams.
Read Trialize’s case study Revolutionizing Clinical Trials: How AI and Automation Reduced Study Build Time from Months to Days, explaining the unique technical capacity of Trialize’s platform.
Now, let’s get to this week’s topics!
Weekly tech+bio highlights
💰 Vertex Pharmaceuticals acquires Alpine Immune Sciences for $4.9 billion, marking the largest biopharma acquisition of the year. This strategic move strengthens Vertex's position in immunology, with plans to further develop Alpine's lead asset, povetacicept, for IgA nephropathy and other autoimmune conditions.
🔬 Ginkgo Bioworks introduces Lab Data as a Service, offering rapid, scalable data generation for biotech R&D. With services like Enzyme Performance and Antibody Developability, customers can access extensive data in weeks, starting at prices like $78,000 for enzymes and $48,000 for antibodies, enhancing machine learning model training and design testing.
💰 e-therapeutics plans to delist from the London Stock Exchange and potentially relist on Nasdaq, following a £29m fundraise to invest in its AI-driven drug discovery platform. Despite strong backing from major shareholders, the UK investor response has been underwhelming, prompting the move to seek better valuation in the US market.
🔬 Ginkgo Bioworks and Novo Nordisk expand their strategic partnership, aiming to enhance the R&D and manufacturing processes for Novo Nordisk's chronic disease medications, including those for diabetes and obesity. The collaboration, set to last at least five years, will focus on leveraging Ginkgo's synthetic biology platform across the entire R&D pipeline.
🔬 Google and Bayer team up to develop AI-driven technologies for radiologists, integrating Google's cloud and AI capabilities with Bayer's expertise in radiology and clinical data. This collaboration will focus on creating a platform to ease radiologists' workloads and accelerate innovation in medical imaging, with plans for a first release in the EU and U.S. later this year.
🔬 Neurotech startup Neurovalens secures FDA clearance for a noninvasive anxiety treatment device, further expanding its portfolio which also includes an FDA-cleared device for insomnia.
💰 ADC-focused German biotech Tubulis secures an upsized $138 million Series B2 funding round, led by EQT Life Sciences and Nextech Invest. The funds will advance Tubulis' pipeline of antibody-drug conjugates targeting solid tumors and support the establishment of a U.S. subsidiary.
🚀NFX launches a $20 million ultra-fast investment track for Israeli AI and biotech startups, offering Seed and pre-Seed funds with a commitment to respond within nine days and transfer funds in three weeks, streamlining investment with SAFE agreements to reduce bureaucracy.
💰 Copenhagen-based startup Reshape Biotech secures $20M in Series A funding to expand its AI and robotics lab automation technologies to the US, enhancing R&D in biotech, agriculture, and food industries. Investors include Y Combinator and Astanor Ventures.
🔬 Children's Tumor Foundation invests in Healx to progress treatments for Neurofibromatosis Type 1, supporting the advancement of Healx's AI-powered drug discovery platform and lead compound, HLX-1502, into Phase 2 trials. This partnership emphasizes a patient-centric approach to address the significant unmet needs of the estimated 3 million global NF1 patients.
🔬 Charles River Laboratories and Deciphex launch Patholytix Foresight, an AI-powered decision support tool for toxicologic pathology, enhancing the efficiency and accuracy of histopathological assessments.
💰 DeepCure secures $24M in Series A-1 funding led by IAG Capital Partners to enhance its AI and automation capabilities in small molecule drug discovery, focusing on advancing its immunology and inflammation programs towards clinical trials
Seed Health Invests in AI-Driven Biology Platform to Enhance Microbiome Science
A 2015-founded life sciences startup, Seed Health has unveiled CODA, a new computational biology platform designed to leverage the expansive Human Phenotype Project data, managed by Professor Eran Segal at the Weizmann Institute of Science.
This initiative aims to foster the development of precision probiotics and other microbiome-directed interventions, enhancing the understanding of the intricate relationships between the microbiome and various health outcomes.
CODA will initially focus on advancing research in areas such as cardiometabolic health, brain health, menopause, and longevity.
The CODA platform integrates over three million phenotypic data points from more than 13,000 individuals, encompassing 40,000 cumulative subject years.
This robust dataset merges microbiome analysis with genetic, immune, metabolomic, and proteomic data, offering a holistic view of human health influences—biological, environmental, dietary, and lifestyle factors, along with medical histories.
This comprehensive approach aims to elucidate the complex interactions between the microbiome and host health.
The CODA platform's depth allows for the analysis of the microbiome’s interaction with various biological systems, such as the gut-brain and gut-heart axes.
Notable achievements include the development of organ-specific clocks that can differentiate between biological and chronological aging across different organs, highlighting the microbiome's potential as a target for health interventions.
Professor Segal’s research has been instrumental in identifying microbial genes and their network effects, which are critical for understanding health and disease mechanisms.
This research has led to discoveries such as the linkage between microbial genes and body composition, as well as the diet’s effect on metabolic parameters through microbiome modulation.
Novel Gen AI Platform for Preclinical Research
I’ve been following Toronto-based BenchSci for years, and I remember they started as an AI-search engine for finding reagents online with a focus on antibodies, including protocols for designing experiments.
And now, years later, the platform has kind of pivoted into a new realm: building a generative AI-powered drug discovery engine called Ascend.
Ascend utilizes multimodal AI to analyze and interpret complex biological data from public and private sources, including text and figures, with a focus on scientific accuracy and explainability.
By integrating and processing diverse datasets securely, Ascend supports preclinical R&D workflows, facilitating better experimental design, target identification, and drug discovery.
This company's ML model predicts KRAS inhibitor's clinical benefit in patients with >80% accuracy
Recently, Rafael Rosengarten pointed me to a research poster by Genialis introducing 𝗚𝗲𝗻𝗶𝗮𝗹𝗶𝘀™ 𝗸𝗿𝗮𝘀𝗜𝗗: a new AI/ML classifier that accurately predicts KRAS outcomes in lung cancer patients.
Most biomarkers in the KRAS space measure only KRAS mutation status, which may be necessary to receive the drug but is insufficient to predict response or to inform further therapeutic interventions.
Genialis krasID works better because it is built differently.
By using high-dimensional gene expression data and a ML algorithm comprising numerous biologic signatures, it captures a more complete view of the underlying biology 𝘂𝗻𝗶𝗾𝘂𝗲 𝘁𝗼 𝗲𝗮𝗰𝗵 𝗶𝗻𝗱𝗶𝘃𝗶𝗱𝘂𝗮𝗹 𝗽𝗮𝘁𝗶𝗲𝗻𝘁.
Because the model incorporates biology aspects that are intrinsic to KRAS function in the tumor and extrinsic to the surrounding tumor milieu, this biomarker is able to perform well for different drugs across KRAS-driven cancers.
The results, presented at hashtag#AACR24, outline performance of Genialis krasID in predicting clinical benefit in a real-world cohort of patients who received sotorasib, a KRAS inhibitor currently approved for KRAS G12C non-small cell lung cancers. 💊
Today’s initial results demonstrate that the Genialis krasID classifier can predict response:
✅ with >𝟵𝟰% 𝗮𝗰𝗰𝘂𝗿𝗮𝗰𝘆 across multiple different KRASi in preclinical models and
✅ with >𝟴𝟬% 𝗮𝗰𝗰𝘂𝗿𝗮𝗰𝘆 in real world patients.
The company announced the launch of an 𝗲𝗮𝗿𝗹𝘆 𝗮𝗰𝗰𝗲𝘀𝘀 𝗽𝗿𝗼𝗴𝗿𝗮𝗺 for Genialis krasID, where members will gain the opportunity to test drive the biomarker model on their data and establish a foundation for implementing the biomarker within their translational workflows.
Model Medicines Advances AI Drug Discovery in Antiviral and Oncology Fields
What? Model Medicines has introduced new methodologies in drug discovery through two preprint papers, demonstrating significant advancements in targeting antiviral and oncology diseases using artificial intelligence.
The first paper focuses on identifying the RdRp Thumb-1 site as a new target for broad-spectrum antivirals and introduces MDL-001, a compound showing promise as an orally bioavailable broad-spectrum antiviral.
The second paper discusses the ChemPrint deep learning model, part of the GALILEO AI platform, which excels in predicting molecular properties and has successfully identified novel inhibitors for previously "undruggable" oncology targets AXL and BRD4.
Why? The development of these AI-driven methodologies is crucial for addressing the challenges in conventional drug discovery processes, which are often slow and have high failure rates. By leveraging AI, Model Medicines aims to enhance the speed and efficiency of discovering novel therapeutics for diseases that have high unmet medical needs, such as various viral infections and cancer. The GALILEO AI platform enables the exploration of deeper chemical spaces and the prediction of promising compounds more accurately, potentially leading to more effective treatments.
Takeaways
Novel Targets and Compounds: The discovery of the RdRp Thumb-1 site and the development of MDL-001 signify a new direction in broad-spectrum antiviral research, showing effective in vitro and in vivo results.
ChemPrint Model Efficiency: The ChemPrint model has outperformed traditional AI approaches in a prospective study, achieving a 45.5% hit rate in identifying inhibitors for challenging oncology targets, thereby expanding the possibilities in drug development against tough diseases.
A Novel Quantum-assisted Approach to Designing Drugs From Fragments
A new algorithm called Quantum-assisted Fragment-based Automated Structure Generator (QFASG) from a team of researchers at clinical stage artificial intelligence (AI)-driven biotech company Insilico Medicine offers a meaningful step forward in automated drug design.
QFASG utilizes advanced chemoinformatic techniques to construct novel small-molecule structures from a library of molecular fragments.
To demonstrate its effectiveness, the researchers successfully applied QFASG to generate new small-molecule inhibitors of two pharmacological targets involved in cancer pathogenesis. The findings were published in Frontiers in Chemistry.
In recent years, AI-assisted generative chemistry platforms such as Insilico’s Chemistry42 engine – a software platform for de novo small molecule design and optimization that integrates AI techniques with computational and medicinal chemistry methodologies – have demonstrated the possibility of designing novel drug candidates from scratch using generative AI. Now, driven by advances in quantum computing and chemoinformatic techniques, researchers can explore vast chemical spaces with new speed and efficiency, allowing them to design new compounds from virtual fragments.
"QFASG generates new drug-like structures using a medicinal chemistry approach," says Sergei Evteev, a lead researcher on the project. "By leveraging molecular fragments and advanced algorithms, we can design compounds that interact specifically with target proteins, potentially leading to detection of new promising chemotypes."
One of the key features of QFASG is its fully automated pipeline, which streamlines the process of structure generation and optimization.
The algorithm employs a two-part approach: first, probes are strategically placed within target binding sites, and then structural growth occurs through iterative modules, resulting in the generation of diverse compound libraries.
"Our goal with QFASG is to provide researchers with a powerful tool for generating primary hit compounds," says Alex Zhavoronkov, PhD, founder and CEO of Insilico Medicine and a member of the research team. "These initial structures can serve as starting points for further optimization and development, ultimately accelerating the drug discovery process."
To demonstrate the effectiveness of QFASG, the researchers used the algorithm to design inhibitors for two proteins that hold promise as targets for new cancer treatments: ATM and CAMKK2 kinases. ATM kinase is involved in DNA damage response and repair and is a promising target for the treatment of solid tumors, glioblastoma, sarcoma, lung, prostate and colorectal cancer and other pathologies. CAMKK2 plays a significant role in various physiological and pathological processes, including cancer cell growth and is overexpressed in multiple tumor types, including prostate, breast, ovarian, gastric and hepatic cancers
In order to test the capabilities of QFASG, the researchers performed insilico validation by reproducing binding poses of known protein ligands and in vitro evaluation of structures generated for ATM and CAMKK2. In both cases, three unique structures designed by the algorithm were selected for synthesis and testing and several compounds demonstrated low micromolar activity against the desired targets.
Early cancer vaccine data breathe new life into the field, but many questions remain: #AACR24
An insightful report from AACR24 by Endpoints News sheds some light on the emerging field of cancer vaccination.
What: At the recent American Association for Cancer Research (AACR) conference in San Diego, a diverse array of experimental cancer vaccines was introduced. These range from "off-the-shelf" options to personalized vaccines designed to align with the specific genetic mutations of a patient's tumor.
Why: The development of cancer vaccines has been pursued for decades, but only recently have technological advancements in mRNA and other genetic delivery systems enabled more effective formulations. These vaccines aim to train the immune system to recognize and fight cancer cells by targeting neoantigens unique to individual tumors.
Takeaway: Early trials, though small, have shown promising results in preventing the recurrence of certain cancers, such as pancreatic and head and neck cancers. This reinvigorated research interest highlights a shift towards using vaccines at strategic points in cancer treatment, potentially offering a more targeted and effective approach to cancer management. The ongoing debate and research are focusing on optimizing the immune response either through customized or off-the-shelf vaccines. This could pave the way for more effective therapies and, ultimately, more hopeful outcomes for patients facing various forms of cancer.
Mapping AI-driven drug candidates and targets (infographic)
Recently, we launched a report, “It’s Been a Decade of AI in the Drug Discovery Race. What’s Next?“ where we profiled the historical progress of the internal drug discovery pipelines of 9 leading AI platforms.
Here, we are continuing this study with this infographic (still in process), where we estimate the novelty of targets for some of the candidates in development:
