Questions about Proteus Bio.

Proteus Bio is building an AI-native pharmaceutical company. Proteus One is the intelligence architecture we are developing to connect target discovery, therapeutic discovery and clinical forecasting across an owned portfolio. We call this operating model neo-pharma. The connected architecture is in development.

At Proteus, biological intelligence means a system that uses evidence about living systems to guide biological research and design. Our aim is to connect reasoning about disease, therapeutic candidates and experimental results so that relevant findings inform the next decision.

We are building drug-discovery capability into a shared AI architecture from the outset. The aim is to coordinate multiple therapeutic campaigns in parallel and retain what each learns across the portfolio. Proteus One connects target discovery, therapeutic discovery and clinical forecasting; that connected architecture is in development.

Proteus Bio is building a pharmaceutical company around an owned portfolio and a shared intelligence architecture. Our work combines therapeutic research, computational design and experimental evidence. Research and partnership enquiries can be directed to our team.

Proteus Bio is based in San Francisco. Our contact page is the starting point for research, partnership and investment enquiries.

Yes. Contact us with the target, biological question and the evidence you want to establish. We can discuss whether it fits our research direction and the work a collaboration would involve.

De novo antibody design means generating new antibody sequences or structures for a specified binding objective. Computational candidates are then made and tested in the lab.

De novo protein design means creating a new protein sequence or structure to meet a desired function, such as binding a target. Experiments establish how well the design meets that objective.

Yes. Computational design can generate new protein or antibody candidates, which can then be tested experimentally. Proteus has reported a first VHH binder benchmark; the target and affinity results are described below.

A design workflow specifies the target, binding objective and constraints, generates candidate sequences or structures, and selects designs for experimental testing. Measurements can then inform further design decisions.

Lab-in-the-loop means experimental observations inform subsequent computational or research decisions. A loop is closed when a result changes what happens next, such as the candidate selected or the experiment designed. This feedback is part of the Proteus One architecture we are developing.

Autonomous experimental research is part of the architecture we are developing. Computational design decisions, laboratory execution and learning from experimental feedback are distinct capabilities that Proteus One is intended to connect. The connected architecture is in development.

Proteus has reported VHH binders against PD-L1, PDGFR, the insulin receptor and TNF-α in its first benchmark, with ten designs tested per target and six binders below 50 nM measured by biolayer interferometry (BLI). These company-reported results predate their use as experimental feedback. They concern target binding, not therapeutic efficacy or validation of the full Proteus One architecture.

A validated binder has experimental evidence that it binds the intended target under the tested conditions. Proteus’s reported benchmark used BLI to measure binding. Specificity, functional activity, developability and clinical benefit are separate questions and require their own evidence.

VHHs, also called nanobodies, are single-domain antibody formats. De novo design generates candidates for a defined objective before experimental testing. Proteus’s first reported binder benchmark used VHH designs.

Structure prediction and protein language models can supply useful capabilities within a research workflow. Proteus One is the broader architecture we are developing to connect target reasoning, therapeutic design and clinical forecasting, using computational tools and experimental evidence to guide campaigns.

Proteus is the therapeutic-discovery role within Proteus One. It is intended to coordinate a campaign using design and evaluation tools. The wider architecture connects that work to target discovery and clinical forecasting.

A design model generates or evaluates molecular candidates. Proteus is intended to coordinate the therapeutic campaign: its objective, tools, candidate choices and experimental evidence.

An AI-native pharmaceutical company builds AI into its core research capability and operating model. Proteus Bio’s neo-pharma model uses a scalable AI architecture to connect target discovery, therapeutic discovery and clinical forecasting across multiplexed campaigns. It is the model we are developing.

For Proteus, it means starting with a biological objective, generating candidates to meet it, and testing what they do.

Techbio and AI-native pharma are broad, overlapping terms. Proteus uses neo-pharma to describe a specific operating model: a pharmaceutical company built around scalable AI, an owned portfolio and experimental learning across multiplexed campaigns.

Neo-pharma is a ground-up pharmaceutical model in which drug-discovery capability is built into a scalable AI system. It connects target discovery, therapeutic discovery and clinical forecasting to run large, multiplexed campaigns across an owned portfolio. Relevant findings inform the next decision across campaigns, allowing discovery capacity to expand without building a separate organisation around every programme.

Proteus One is the architecture we are developing for neo-pharma. Asclepius focuses on target discovery, Proteus on therapeutic discovery and Pythia on clinical forecasting. Shared objectives and experimental evidence are intended to connect decisions across multiplexed campaigns. The connected architecture is in development.

Multiplexing coordinates therapeutic campaigns at different stages through a shared intelligence. While one experiment runs, another campaign can design candidates and another can interpret results. The aim is to expand discovery capacity without building a separate organisation around every programme.

Proteus One’s connected architecture is in development. Our first reported binder benchmark concerns the therapeutic-design component and predates those results feeding back into the system. We are building the connections between target discovery, therapeutic discovery and clinical forecasting.

An experimental finding can change the next candidate design, the next measurement or the original target hypothesis. Proteus One is designed to retain relevant findings so they can inform decisions across campaigns.

N-1 therapeutics means working towards treatments designed around an individual’s biology. It is a long-term research direction for Proteus, alongside undruggable targets and rare diseases.

Proteus-SD describes our biosecurity research direction, exploring how biological design could contribute to detection, diagnostics and countermeasures.

Biological design requires attention to both beneficial applications and possible misuse. Proteus-SD explores how that design capability could contribute to detection, diagnostics and countermeasures.