CELL IDENTITY × EXPRESSION TIMECan an immune-cell population be programd in vivo for a defined period?
SCIENCE
Three questions.
One translational thesis.
Byterna studies how RNA format, cell-selective delivery, computational learning and manufacturing can work together to make in vivo immune-cell programming more controllable—and to turn patient-specific biology into testable medicines.
Explore the science ↓TUMOR INFORMATION × ANTIGEN COMBINATIONCan patient-specific mutations become a testable, manufacturable vaccine design?
PREDICTION × EXPERIMENT × PROCESSCan computation make biological uncertainty more tractable?
01 · IN VIVO IMMUNE-CELL PROGRAMMING
Program the right cells—
for a defined period.
Moving CAR-T programming into the body creates two linked scientific questions: which immune-cell population receives the payload, and how long the encoded program remains active.
Byterna combines cell-selective CellectLNP delivery with non-integrating cmCAR circular mRNA. The approach is designed to support time-limited immune-cell programming without permanent genomic modification. BR101 has a registered investigator-initiated trial (IIT) in relapsed/refractory multiple myeloma (NCT07537049). Autoimmune diseases are a planned expansion direction.
See BR101 ↗WHERECell identityCell-selective delivery
WHENExpression windowNon-integrating circular mRNA
IN SITUImmune-cell
reprogrammingDesigned for time-limited, non-integrating expression
reprogrammingDesigned for time-limited, non-integrating expression
CLINICAL QUESTION
Can patient-specific immunity help address residual disease, recurrence risk and tumor evolution?
02 · PERSONALIZED CANCER IMMUNITY
Turn each tumor's mutations into a vaccine hypothesis.
Neoantigens arise from tumor-specific mutations and differ from patient to patient. The scientific challenge is not simply to find mutations, but to identify which candidates are most likely to be processed, presented and recognized by the immune system.
Byterna is developing an AI-guided workflow that connects tumor-normal sequencing and HLA context with neoantigen prediction, ranking and combination optimization. The selected antigen set is carried into circular mRNA construct design, manufacturing and immune testing through a traceable development record.
See the cancer-vaccine program ↗03 · COMPUTATIONAL SCIENCE
Make biological uncertainty testable.
Neoantigen combinations, RNA constructs and manufacturing parameters form a high-dimensional design problem. Deep learning, specialized agents and emerging quantum-computing methods can help generate candidates, prioritize experiments and reveal relationships that are difficult to explore one variable at a time.
The objective is not autonomous drug invention. It is a faster, more traceable scientific cycle in which predictions meet experiments, process constraints enter design decisions and experts review every development gate.
Explore our technology system ↗SHARED EVIDENCEScientific
learning loopExpert review at every gate
learning loopExpert review at every gate
01PREDICTGenerate hypotheses
02DESIGNSelect an experiment
03TESTMeasure biology
04LEARNUpdate the next decision
SELECTED PEER-REVIEWED WORK
Scientific claims should be
open to scrutiny.
COMPUTATIONAL METHODS · RNA DESIGN · 2025A Hyperbolic Discrete Diffusion 3D RNA Inverse Folding Model for Functional RNA DesignJournal of Chemical Information and Modeling ↗PRECLINICAL · EX VIVO T-CELL ENGINEERING · 2025Scarless circular mRNA-based CAR-T cell therapy elicits superior antitumor efficacyPreclinical study using ex vivo-engineered T cells; it did not evaluate CellectLNP-mediated in vivo delivery or report BR101 clinical data · Signal Transduction and Targeted Therapy ↗View publications ↗