AI-designed zinc finger proteins

Reach any gene.
Anywhere in the genome.

TBG Therapeutics has built a powerful AI platform for designing zinc finger proteins — fully human, programmable DNA-binding domains that act as molecular beacons, homing in on any unique sequence in the genome with precision and selectivity. We are opening this platform to biotech and pharma partners who want to reach targets that other technologies cannot.

Partner with us How it works
The TBG platform

AI that reads the genome — then writes the protein to match.

We have trained our AI model on billions of experimental protein–DNA interactions. The result is a closed-loop system that instantaneously designs, tests, and refines zinc finger arrays for any genomic target in minutes.

01
AI-guided design in real time

Input any DNA target sequence. Our model instantly predicts the optimal zinc finger array configuration — trained on 94 billion experimental protein–DNA and protein–protein interactions from 23 curated in vitro screens — narrowing down to the protein sequences most likely to bind your target with high specificity.

02
Optimization for any genomic objective

Once candidate arrays are identified, we optimize constructs for the specific molecular outcome required — whether the goal is transcriptional silencing, gene activation, epigenetic remodeling, sequence editing, or site-specific integration. Optimization strategy is tailored to the biology of the target, not a fixed protocol.

03
Self-improving data flywheel

Every experimental cycle generates new protein–DNA interaction data that feeds back into the AI model, continuously expanding its coverage and sharpening its predictions. The platform compounds with each partner program — a true flywheel, not a static tool.

The biology behind the platform

The largest family of DNA-binding proteins in the human genome.

700+
Zinc finger-coding genesin the human genome, many involved in regulating gene expression
~10,000
Individual DNA-binding domainsacross all zinc finger proteins in the genome
The zinc finger domain is one of the most abundant structural motifs in eukaryotic genomes — present in over 700 human proteins carrying nearly 10,000 individual DNA-binding domains, the great majority involved in regulating gene expression. Because they are native human proteins, the immune system recognizes them as self, giving them a fundamentally diminished immunogenicity profile compared to bacterial or viral-derived editing tools. And because they are modular by nature, they can be re-engineered as targeting scaffolds for virtually any molecular payload.
Why zinc fingers

Precise. Programmable. Delivery-agnostic.

Diminished immunogenicity — a fully human scaffold

Compact — fits any delivery vehicle

Tunable affinity for precise selectivity

Works with every genome engineering modality

Multiplexing for combinatorial programs

Partnership models

Ways to build with TBG.

We structure partnerships around specific genetic targets — preserving your exclusivity while keeping the platform available to multiple collaborators across non-overlapping programs. Each engagement is a genuine scientific collaboration, not a fee-for-service screen.

Target-based discovery

ZFP Discovery & Licensing

You bring the target gene and the therapeutic hypothesis. We design, synthesize, and optimize zinc finger arrays for your genomic locus — delivering a ranked, fully characterized construct panel alongside a platform access license for ongoing design against that target. Exclusivity and scope are defined around the specific genetic target, enabling multiple partnerships in parallel without overlap.

Co-development

Lead Program Co-development

TBG is actively developing internal programs in CNS and cardiovascular disorders — disease areas where the precision, allele-selectivity, and delivery flexibility of zinc finger technology offer meaningful advantages over existing approaches. We are open to partnership discussions with organizations that share a scientific interest in these areas, whether to co-develop an existing TBG program, contribute complementary capabilities, or access our platform data to accelerate a parallel effort.

Modality extension

Payload-Agnostic Targeting

Our ZFP arrays are modular DNA-homing signals compatible with any molecular payload. Partners working on gene insertion, prime editing, chromatin restructuring, recombinase-mediated integration, or novel epigenetic effectors can license TBG’s targeting technology to precisely deliver their payload to any genomic site.

Custom engagement

Don’t see your model here?

We are open to structuring partnerships around the specific needs of your program. If your indication, modality, or collaboration structure doesn’t fit neatly into one of the above, reach out and we’ll design an agreement together.

Scientific leadership

Founders who pioneered the field.

Marcus B. Noyes

Marcus B. Noyes, PhD

Co-founder & CEO · Professor

Pioneer in zinc finger biology since 2002. Expert in transcription factor function with seminal publications spanning ZF domain engineering, computational biology, and genomics. His lab developed many of the experimental approaches that generated the training data underlying the TBG AI platform.

University of Pittsburgh
Philip M. Kim

Philip M. Kim, PhD

Co-founder · Professor

Expert in protein engineering and a pioneer in ML/AI methods for protein design. 80+ publications in computational biology, protein interactions, inhibitor design, and high-throughput screening. Instrumental in developing the machine learning architecture underpinning the TBG platform.

University of Toronto
Mikko Taipale

Mikko Taipale, PhD

Co-founder · Professor

Expert in transcription factor regulatory networks and human regulatory domain biology. 80+ publications in functional proteomics, protein–protein interactions, and systems biology, with a focus on rare diseases and novel technology development.

University of Toronto
Get in touch

Start a partnership conversation.

We work with a focused set of biotech and pharma partners at any given time. If your program involves a genetically defined target where ZFP technology could add value, we’d like to hear from you.

Send inquiry