Abstract
In this interview, Dr. Christian Rohlff, CEO of Oxford BioTherapeutics, discusses how shifting from genomics to AI-driven proteomics can break the biotech industry’s echo chamber of redundant drug development. He further outlines the strategic partnerships and machine learning innovations required to reach the significant portion of oncology patients who are underserved by current therapeutics.
Overview
Founded in 2004, Oxford BioTherapeutics (OBT) has quietly built one of the most robust proteomics-driven oncology pipelines in the industry. On behalf of MedTech Soc, I sat down with founder and CEO Dr. Christian Rohlff to discuss his inspirations and OBT’s specialised technique for target identification. We also explored the strategic decisions that propelled this Oxford-based startup to approximately $150 million in revenue, achieving remarkable growth amidst a prolonged period of industry-wide capital scarcity. Projected to reach $1 billion in revenue within the coming decade, and supported by partnerships with several pharmaceutical giants like Roche, Boehringer Ingelheim and GSK, OBT currently has around 20 molecules in joint R&D. Furthermore, with dozens of additional assets still available for other partners, OBT presents a masterfully executed commercialisation structure designed to maximise clinical impact.
The Introduction: Dr. Rohlff, Academia, and Industry
Dr Rohlff got his start in academia in Germany before moving to the US to complete his PhD at Georgetown University, ultimately ending up in the academic labs of the National Cancer Institute (NCI). He describes a meaningful shift in his perspective while working at the NCI, where the research labs were situated right across the hall from the patient wards. The proximity to patient care allowed him to interact with clinicians and witness firsthand the consequences of cancer.
Although Dr Rohlff completed his PhD in Pharmacology with the intention of remaining in academia, an industry grant enticed him to explore biotechnology and brought him to Oxford where he subsequently founded his company. This exposure came with the realisation that if he wanted to bring highly applied, translational research directly to the patient, biotech was the ideal path as it offered a much faster route to advance candidate medicines to patients.
Dr Rohlff: It brings me back to the start of my career when I was at the National Cancer Institute where our labs were right next door to the clinical facilities. I am reminded very starkly of the feeling of knowing that the clinicians only had a limited number of options they could offer a patient. Beyond that there’s nothing else they can do and the clinician cannot really change that. We as scientists can change that. We can give them additional options. And that’s what I call the moral imperative, I think that’s at the core mission of what the company wants to do.
For Dr Rohlff, this imperative is profoundly personal and grounds the company’s scientific output in an understanding of patient vulnerability. The early loss of his sister to the disease inspired a resolve that sustained him through the industry’s hurdles. Later, his own cancer diagnosis reinforced this perspective by exposing him to the reality of surrendering one’s medical autonomy to an oncologist. Informed by this dual vantage point as an industry leader and as a patient, Dr Rohlff has structured OBT as a highly pragmatic enterprise. Today, the organisation functions with the operational directive of creating and accelerating highly effective, clinical-stage therapeutics.
The Science: Proteomics, Target Identification, and OGAP®
Historically, target selection for cancer drugs has relied on genomic and transcriptomic approaches. In these models, researchers sequence the genes of the cancerous tissue to identify mutations, operating on the assumption that these abnormalities will reliably translate into targetable proteins on the cell surface2,3. While this approach has yielded significant clinical success, it inherently biases target discovery toward previously characterised proteins. Systemic pressures within academia often discourage the investigation of novel targets, with funding bodies penalising grant proposals focused on unstudied proteins, where the clinical rationale is harder to prove. Furthermore, the demand for rapid results naturally incentivises scientists to rely on established biological systems where research infrastructure and reagents are already available. Over time, these academic bottlenecks, amplified by bias among venture capitalists and major pharmaceutical companies seeking safe financial bets, have created an echo chamber of redundant development5. Consequently, biotech companies are left researching the same small fraction of the proteome, leading to the crowded pursuit of “validated” targets like HER2, TROP2, and Claudin 18.24. This exacerbates a clinical landscape where the popular targets, despite already having many approved products and more in clinical trials, continue to receive the majority of R&D funding, even though novel targets would likely offer better efficacy with significantly less off-target toxicity4,6.
Recognising that genomic and transcriptomic target validation is resource-intensive and frequently constrained by the factors above, Dr Rohlff adopted a target identification approach focusing on proteomics. Using mass spectrometry analysis, OBT analyses the cell surface protein content directly from patient tissue biopsies rather than inferring membrane composition from mRNA transcripts, allowing for faster and more cost-effective target identification.
The infrastructure behind this methodology is OGAP®-Verify, which is the world’s largest quantitative membrane protein expression database, containing data on ~7,000 membrane proteins across most known malignancies as well as normal tissues. Dr Rohlff purchased this second-generation database at OBT’s inception and today it is capable of detecting proteins with just 50 copies per cell, far surpassing the capabilities of classic immunohistochemistry. This enhancement constitutes what Dr Rohlff calls “a major milestone in the history of the company.” By pinpointing cancer-specific targets which are virtually absent on normal tissues, OBT minimises the risk of off-target toxicity and perfectly positions the company to develop bespoke new drugs in the forms of antibody-drug conjugates (ADCs) and T-cell engagers.
Fiona: 2025 was a big year for OBT, particularly in securing the Roche and GSK partnerships. Why has the current landscape of ADC development suddenly made Pharma so excited to get involved in projects with proteomics? What makes your platform a differentiator?
Dr Rohlff: Innovation comes in waves. A decade ago, the big innovation was immune checkpoint inhibitors. That was a tremendous milestone, and with those antibodies, clinicians could for the first time cure patients. But only 20% of patients would have a durable response. For the other 80%, ADCs and T-cell engagers are really at the forefront of the next wave of innovation in oncology. We are tearing up old rule books and making new rules because things are unexpected. We are seeing greater efficacy and dramatically increased safety profiles for ADCs. The main limitation is that the currently pursued drug targets for ADCs only cover about 20% of the population who need them. For the other 80%, we need other receptors with similar characteristics to address that unmet need. That’s really the big gap that we at OBT fill. In conjunction with having a unique, differentiated platform technology that demonstrates how you can pursue this novel patient population with a high predictive value of success, that is where we add value.
To further address this gap, OBT is developing OGAP®-Splice, an AI-enabled platform scheduled for release within the year focusing on oncologically-relevant splice variants. While the human genome contains approximately 22,000 genes, alternative RNA splicing generates roughly five times as many distinct protein variants. Through its proteomic analysis of human tissue samples, OBT isolated ~100,000 peptide sequences that did not map to canonical genetic structures. To identify these peptides, the company constructed a dataset of splice variants, or a spliceosome, of the human genome, and mapped these sequences to known protein structures. The pilot study which validated this approach revealed targets with unprecedented tumour selectivity.
Dr Rohlff: We identified a splice variant of a canonical gene where the consequence of the variation was that a typically cytosolic protein now contained a plasma membrane domain. This variation relocated the protein onto the cell surface exclusively in tumour cells. In normal tissue, the protein remained intracellular. Because of this differential localisation, a therapeutic antibody can target the tumour-specific surface variant without binding to the normal tissue, where the protein remains inaccessible inside the cell.
This specific variant is highly prevalent, presenting in up to 80% of certain tumour types. OBT subsequently completed the preclinical validation and partnered the asset with a major pharmaceutical company, which is currently developing a targeted therapeutic antibody.
While this pilot study proved that splice variants can yield highly selective targets, scaling this discovery process requires advanced computational infrastructure. Integrating machine learning into the OGAP®-Splice platform has therefore proven transformative in terms of both computational efficiency and high-throughput discovery.
Dr Rohlff: We have manually annotated variants in the past, but the analysis is highly complex. This is a prime example of where AI and machine learning are uniquely powerful. We possess a massive, proprietary dataset to train the system, allowing us to compute and validate novel splice variant targets at a scale that would traditionally take years of manual research.
However, while acknowledging the transformative potential of machine learning in target discovery, Dr. Rohlff cautions against overestimating these computational models. He emphasises that a model’s utility is strictly context-dependent, relying heavily on the quality and availability of data specific to the field of study.
The Business: Markets, Milestones, and Partnerships
The biotech sector has seen intense volatility in recent years. Driven by historically low interest rates and investor enthusiasm from the pandemic, the industry witnessed an unprecedented influx of capital between 2020 and 2021. This surplus of funding temporarily altered the traditional drug development paradigm. Rather than licensing validated assets to major pharmaceutical partners after Phase I studies, many early-stage firms used this money to independently finance Phase II and III clinical trials for the first time. However, when the pendulum swung backwards, interest rates rose and the capital markets abruptly contracted. Without follow-on financing, numerous companies burdened by the massive operational burn rates of clinical trials were forced into pipeline reprioritisation, mass layoffs, or complete insolvency. This contraction resulted in a dramatic decline in IPOs and IPO valuations. Consequently, a dramatic reversal in investor confidence precipitated a period of severe capital scarcity and low venture investment throughout 2023 and 20247,9.
Entering the 2025 and 2026 financial cycles, Dr Rohlff reports the sector seems to be returning to realignment. After biotech firms carry out the early Phase I trials and pass their de-risking milestones, assets are handed over through licensing or acquisitions to pharma companies, who possess the massive infrastructure required to handle late-stage trials and commercialisation. Currently, roughly two-thirds of major pharmaceutical pipelines are sourced externally from these early-stage partnerships.
Throughout the funding cycles of recent years, OBT adhered to this traditional model. By focusing exclusively on target identification and early clinical development, OBT avoided some of the difficulties that crippled other early-stage firms. As Dr Rohlff noted, this symbiotic approach allows biotech firms to play to their strengths in innovation while leveraging the critical mass and infrastructural expertise of pharma giants to successfully bring therapies to market.
Fiona: Do you see a situation in which you ever do a wholly owned pipeline to take a drug from initial research all the way through to commercialisation, or are you happy with the strategic partnership model?
Dr Rohlff: It’s the latter. If I were to pick one of our drug targets, make a therapy for it, and fund it all the way to the market, it’s highly unlikely that I would be able to do more than one molecule at a time. The potential impact from that perspective is quite low. With the model we have now, partnering with Big Pharma companies, we have about 20 molecules that we are researching and developing. We plan to more than double that number in the next decade. If 15 or 20 of those make it to the market, that impact will be huge. By taking this approach, I can benefit a lot more patients.
This strategic reliance on major partners optimises OBT’s strengths and patient impact, but Dr Rohlff is candid about where the true financial gain lies within the pipeline, especially with the advent of AI.
Fiona: Looking ahead to the next decade, who do you believe will capture the most value in this sector? Will it be the tech platform companies that own the data and train the models, or will the traditional pharmaceutical giants maintain an advantage through the capital required to bring the discoveries to market?
Dr Rohlff: “If you look at the incremental steps in value creation, it starts with the discovery aspect which I think to a large extent has been commoditised… 20 years ago, we saw large licensing deals from big pharma companies just to get access to the fundamental technologies needed to create therapeutic antibodies. Today those are so commoditised you can buy them off the shelf for a dime.”
“…the biggest incremental jump in value still happens in the transition between preclinical research and clinical research. It’s all well and fine to test molecules to cure mice from cancer. But going to humans is a completely different story… it’s a big, if you wish, emotional hurdle to actually demonstrate it’s safe in humans.”
“…to answer your question, I would say the ultimate benefactor will still be the pharma company, even if it is the biotech companies who are actually the ones applying AI to create the molecule.”
TLDR: The Science, Business, and Oncology
Oxford BioTherapeutics exemplifies a shift from traditional genomic-based target discovery toward a high-throughput, proteomic model. By leveraging the OGAP®-Verify database to identify membrane proteins and tumour-specific splice variants directly from patient tissues, the company bypasses the validated target echo chamber that often leads to redundant R&D. This technical differentiation, now being scaled through the AI-enabled OGAP®-Splice platform, addresses the 80% of the oncology population currently underserved by existing antibody-drug conjugates (ADCs) and T-cell engagers.
Commercially, OBT’s success involves adherence to a strategic partnership model. By focusing on early-stage discovery and Phase I de-risking, the company has protected itself from the high operational burn rates that led to the end of many biotech peers during the recent capital contraction. The integration of proprietary datasets with machine learning reflects a pragmatic adoption of AI to accelerate discovery while maintaining human oversight to ensure data integrity. As OBT moves toward its goal of $1 billion in revenue, its trajectory highlights a sustainable blueprint for biotech innovation where clinical impact is maximised through technical precision and optimised commercial structures.
References
1. Edwards, A.M., Isserlin, R., Bader, G.D., Campagne, F., Conklin, B.R., Mengod, G., Munos, B., Oprea, T.I. and Schapira, M., 2011. Too many roads not taken. Nature, 470(7333), pp.163-165. Available at: https://www.nature.com/articles/470163a.
2. Urban, S., 2016. SnapShot: Cartography of Intramembrane Proteolysis. Cell, 167(7), p.1898. Available at: https://www.cell.com/cell/fulltext/S0092-8674(16)31729-9.
3. Jackson, D., 2024. ADC Target Selection Summit Meeting 2024. Jackson Consulting Group. [Blog] 21 December. Available at: https://www.jackson-consulting-group.com/post/adc-target-selection-summit-meeting-2024
4. Zhang, X., Huang, P., Jiang, T., Wu, A. and Ding, Y., 2024. Development of antibody-drug conjugates in cancer: Overview and prospects. Cancer Communications, 44(2), pp.255-258. Available at: https://pubmed.ncbi.nlm.nih.gov/38159059/.
5. Barnscher, S.D., 2023. The Clinical Landscape Of ADCs In 2023: Diverse Technologies, Narrow Target. Clinical Leader. [online] Available at: https://www.clinicalleader.com/doc/the-clinical-landscape-of-adcs-in-diverse-technologies-narrow-target-0001
6. Oprea, T.I., et al. 2018. Unexplored therapeutic opportunities in the human genome. Nature Reviews Drug Discovery, 17(5), pp.317-332. Available at: https://pmc.ncbi.nlm.nih.gov/articles/PMC6339563/.
7. Senior, M., 2024. Biotech financing: darkest before the dawn. Nature Biotechnology, 42, pp.1331–1338. Available at: https://www.nature.com/articles/s41587-024-02357-2.
8. Golec, J. and Vernon, J.A., 2009. Financial risk of the biotechnology industry versus the pharmaceutical industry. Applied Health Economics and Health Policy, 7(3), pp.155-165. Available at: https://pubmed.ncbi.nlm.nih.gov/19799470/.
9. IQVIA Institute for Human Data Science, 2023. Global Trends in R&D 2023: Activity, Productivity, and Enablers. [online] Available at: https://www.iqvia.com/insights/the-iqvia-institute/reports-and-publications/reports/global-trends-in-r-and-d-2023
Leave a Reply