Big Picture Bio launches with £2.2m to develop AI-designed cancer combination therapies

London-based biotech Big Picture Bio has launched with £2.2 million in funding to develop AI-designed cancer combination therapies, combining computational modelling with targeted laboratory validation to identify drug combinations with potential clinical impact.

The company has raised £1.5 million in pre-seed funding co-led by Kadmos Capital and Exceptional Ventures, alongside participation from Gloucester Ventures and angel investor John White. A further £700,000 has been awarded through Innovate UK’s Investor Partnerships Programme, bringing total funding to £2.2 million.

Using AI to design cancer drug combinations

Big Picture Bio was founded by Dr Kerstin Papenfuss, CEO and co-founder, and Dr Mark Hammond, CTO and co-founder. Both previously worked with Deep Science Ventures (DSV), where Papenfuss served as Director of Pharma and Hammond co-founded the organisation and led engineering in agentic scientific discovery.

The company is developing a generative world model designed to model how tumours, immune cells and surrounding tissue interact across different diseases and patient subgroups.

Rather than focusing solely on identifying individual drug targets, Big Picture Bio is applying AI to the challenge of cancer combination therapy, where researchers must consider which drugs to use together, as well as factors such as dose, treatment sequence and patient population.

The company aims to narrow this search computationally before moving promising combinations into laboratory testing. Its initial programmes are focused largely on existing medicines and drugs already in clinical development.

From AI modelling to laboratory validation

Big Picture Bio says its platform combines AI-driven modelling with targeted wet-lab experiments. The company plans to use its own experimental data to address gaps in existing datasets and progressively strengthen its models.

The system draws on sources including single-cell and proteomics datasets, CRISPR maps, scientific literature and clinical trial results. According to the company, its models are built around human-defined principles relating to cancer biology, the immune system, drug resistance and drug delivery.

A key feature of the platform is explainability. Rather than producing predictions solely as numerical scores, the company says its system generates a causal chain intended to show how a prediction follows from the underlying biology.

This approach is designed to allow predictions to be reviewed and challenged by scientists, rather than treating AI outputs as a black box.

Testing predictions against clinical outcomes

Big Picture Bio says its platform has been tested retrospectively and prospectively against real-world clinical outcomes, including trial results presented at the American Society of Clinical Oncology (ASCO).

The company says its system correctly predicted the outcome of Regeneron's fianlimab trial in melanoma, including that the tested doses would not meet the study's primary endpoint. European Biotechnology reported that the company had prospectively predicted five of six selected ASCO trial readouts correctly, with an overall prospective accuracy of around 86%, according to the founders.

However, the company says predicting clinical trial outcomes is not its end goal. Instead, it is using these tests to assess whether its models can identify biological relationships that could ultimately inform new treatment combinations.

Moving cancer combinations into the lab

With the new funding, Big Picture Bio says its first AI-designed drug combinations are now moving from computational modelling into wet-lab validation.

The company's initial focus is on solid tumours, where it says extensive single-cell datasets are available. It also plans to investigate existing drugs and previously deprioritised assets, with the potential to develop new intellectual property around drug combinations, dosing, treatment sequencing and patient selection.

The launch reflects a broader push to apply AI and mechanistic modelling to one of the longstanding challenges in oncology drug discovery: determining which combinations of treatments are most likely to work in particular patients.

For Big Picture Bio, the next stage will be to establish whether its computationally designed combinations can translate into reproducible biological effects in the laboratory and ultimately provide a route towards new cancer therapies.