Luminous Ventures is very excited to welcome Phenomic to our portfolio. Phenomic AI, headquartered in Toronto, uses AI/ML to process data emerging from drug screens in complex experimental models that more closely match human biology. This enables Phenomic to identify and develop drugs against novel, disease causing interactions between different cell-types that previously have been difficult to identify using traditional approaches. Luminous Ventures exists to back visionary entrepreneurs and scientists working to solve the world’s biggest problems. Phenomic have now demonstrated that their platform is able to identify new targets involved in cell-cell interactions that enable cancer to resist today’s most powerful medicines. Where many Biotech AI companies are focused on accelerating drug discovery, Phenomic uses AI to open up new areas of biology and ensure targets are relevant to human biology early on. Together, this will enable them to reduce failure rates in the clinic and develop novel high-value medicines that will ultimately improve patient lives.

We at Luminous believe that biological complexity underlies the high rate of failure seen in the clinics. As the drugs targeting easily-modulated mechanisms have been discovered, we are left with more difficult drug targets buried in complex pathways. ‘Basic-research–brute-force bias’ refers to overestimating the ability of advances in basic research, especially brute force screening methods of large compound libraries. Since the 1990s drug discovery efforts have favoured a ‘reductionistic’ approach – screening molecules based on how well they bind to a target protein. This approach can lead to drug candidates that bind tightly to its target protein yet still often fail in clinical trials due to an under-appreciation of how a drug candidate acts in the human body as a whole (Horvath et al. 2016).

There is the need to discover drug candidates using more holistic models that capture the complexity in disease mechanisms (Moffat, Rudolph, and Bailey 2014). Many diseases are driven by multiple cell types interacting with each other within a complex microenvironment. The interactions between different cell types in disease states carry significant information. The tumour stroma has been increasingly recognised as a complex barrier, composed of many different cell types and other factors, such as fibroblasts and the extracellular matrix, that prevent immune therapies from working effectively (Sahai et al. 2020). A major challenge in discovering drugs for the tumour stroma has been screening targets against more physiologically relevant models, such as organoids that incorporate the many cell-types needed to simulate the stroma. Existing in vitro models and analysis methods often remain one dimensional.

Phenomic uses deep-learning tools for image and genetic analysis to deconvolute effects of target inhibition in multi-cell assays. By factoring in multi-omics analysis of tissue data and reverse translation of clinical datasets, Phenomic’s approach seeks to ensure model and mechanism relevance early in the drug discovery process. Its high-throughput analysis capability leveraging AI/ML allows unbiased identification of novel mechanisms and can more accurately capture human disease biology, reducing attrition of candidates.

Phenomic has made impressive progress in perfecting their platform. Building on their foundational machine-learning tools, the team can now process and extract information from the interactions between different cell types at the scale and speed necessary to power drug discovery. They have rapidly identified novel targets in the tumour stroma and have confirmed that they are heavily upregulated in specific solid tumours. Phenomic team are advancing two validated cancer drug targets discovered with their platform into preclinical studies as well as uncovering additional drug targets and building out their candidate pipeline.

Luminous has been closely monitoring the AI-for-drug-discovery space for some time. We have seen that the most ambitious companies in this space are building full-stack suites comprised of technologies such as high content imaging, deep learning, CRISPR, single-cell transcriptomics and automation. This allows companies to address complex biology and to advance their own in-house programmes. Recursion recently announced a $239 million Series D round. Insitro has raised $200 million since its inception in 2018. Both companies are leveraging AI/ML to generate vast datasets and advance their pipelines in-house.

AI/ML tools need to integrate well into the drug discovery and development process to maximise their potential. We are very excited that Phenomic are joined by significant expertise in drug discovery and development. Dr. Michael (Mike) Briskin is appointed as Chief Scientific Officer (CSO); Nobel Laureate James (Jim) P. Allison, Ph.D. and Padmanee (Pam) Sharma, M.D., Ph.D. are appointed to the company’s Scientific Advisory Board (SAB). Dr. Briskin has more than 25 years of industry experience, including founding and running discovery research at Jounce Therapeutics, leading translational studies for inflammation and oncology at  Merrimack Pharmaceuticals, and performing early discovery work for Entyvio® (vedolizumab) at LeukoSite (acquired by Millennium Pharmaceuticals). He currently serves as chairperson of the SAB of Obsidian Therapeutics.

Dr. Allison was awarded the Nobel Prize in 2018 for his pioneering and unrelenting research in cancer immunotherapy, which led to the U.S. FDA approval in 2011 of ipilimumab, a significant life-extending therapy for late-stage metastatic melanoma. His current work is focused on identifying new targets to further unleash the immune system and eradicate cancer. Dr. Sharma is a leader in immuno-oncology clinical trials and drug development, with translational research focused on identifying mechanisms of response and resistance to immunotherapy. Dr. Sharma is currently exploring combinations of immunological therapies and other drugs in preclinical models as well as early-stage clinical trials to potentially overcome specific resistance pathways and improve clinical outcomes for cancer patients.

The appointments of Mike, Jim, and Pam, leaders in immuno-oncology and drug development further highlight the exceptional advancements of Phenomic. Their deep expertise will provide tremendous insights into the biology and clinical strategies for the company’s targets going forward.

We are very proud to support the amazing team at Phenomic on their journey and also to be part of a fantastic investor syndicate including CTI Life Sciences, AV8 Ventures, Viva BioInnovator, as well as existing investors Garage Capital, Hike Ventures, and Cantos Ventures.

Horvath, Peter, Nathalie Aulner, Marc Bickle, Anthony M. Davies, Elaine Del Nery, Daniel Ebner, Maria C. Montoya, et al. 2016. “Screening out Irrelevant Cell-Based Models of Disease.” Nature Reviews. Drug Discovery 15 (11): 751–69.

Moffat, John G., Joachim Rudolph, and David Bailey. 2014. “Phenotypic Screening in Cancer Drug Discovery—past, Present and Future.” Nature Reviews. Drug Discovery 13 (8): 588–602.

Sahai, Erik, Igor Astsaturov, Edna Cukierman, David G. DeNardo, Mikala Egeblad, Ronald M. Evans, Douglas Fearon, et al. 2020. “A Framework for Advancing Our Understanding of Cancer-Associated Fibroblasts.” Nature Reviews. Cancer 20 (3): 174–86.