Cancer vaccine for Yash, he's diagnosed with stage IV pancreatic cancer

The repository documents an effort by the RNA Foundry to develop a personalized mRNA cancer-vaccine design for Yashdeep Bindal, diagnosed with stage IV pancreatic cancer in January 2025. Its lead, Design #2, uses tumour DNA and RNA alongside matched-normal white-blood-cell DNA to identify candidate targets while excluding mutations also found in normal tissue. The design combines 12 mutation regions, five splice regions, five cancer-testis-antigen or self-antigen regions, and two fusion regions, assembled into a 724-amino-acid protein. The repository also preserves Design #1, an earlier, tumour-only design with 29 selected regions, for comparison.

The project shares its selection methods, evidence, code, and instructions so researchers can verify the lead design or propose alternatives. Users can locally check recorded evidence and reproduce selection, antigen ordering, linker optimization, and synonymous sequence encoding; the replay compares outputs byte for byte but does not reprocess raw sequencing reads or run new model inference. Some data are available through a separate read-only endpoint, while original normal FASTQs and complete upstream RNA-processing modules are not included. The design’s documentation records limitations, including sparse RNA support, uncertainty about some RNA evidence, and an unmet soft folding target. It is a research design, not evidence of a clinically tested or effective treatment, and the maintainers invite contributions exploring other antigens, prediction methods, and vaccine architectures.