Eryney Marrogi

Though I’ve spent a good chunk of my professional career in industry, what follows focuses mainly on my academic work.

Cancer

I got an early start in science as a high school student working with Miriam Poirier and Jing Huang at the National Cancer Institute. In addition to falling in love with biology and medicine, I got to contribute to basic biology projects focused on determining the carcinogenic effects of the HIV drug AZT, and later on specific genes involved in osteosarcoma.

• Perinatal exposure of patas monkeys to antiretroviral nucleoside reverse-transcriptase inhibitors induces genotoxicity persistent for up to 3 years of age. The Journal of Infectious Diseases, 2013.
• A RUNX2-mediated epigenetic regulation of the survival of p53 defective cancer cells. PLOS Genetics, 2016.
• cFOS-SOX9 axis reprograms bone marrow-derived mesenchymal stem cells into chondroblastic osteosarcoma. Stem Cell Reports, 2017.

Mosquitoes

As an undergraduate student at Virginia Tech, I had the privilege of working in George Church’s lab on engineering mosquitoes with gene drives to eradicate malaria. The work ultimately did not pan out in achieving our goal, a sustainable gene drive that can continuously propagate, but we did build a population suppression system.

• CRISPR-mediated germline mutagenesis for genetic sterilization of Anopheles gambiae males. Scientific Reports, 2024.
• Engineering gene drive docking sites in a haplolethal locus in Anopheles gambiae. Scientific Reports, 2025.

As part of this project, I also worked alongside mosquito expert Flaminia Catteruccia at Harvard. It’s here that I ironically got my first exposure to machine learning by building an image detection model. I additionally contributed to work investigating basic mosquito biology.

• OocystMeter, a machine-learning algorithm to count and measure Plasmodium oocysts, reveals clustering patterns in the Anopheles midgut. bioRxiv, 2025 (preprint).
• A mating-induced reproductive gene promotes Anopheles tolerance to Plasmodium falciparum infection. PLOS Pathogens, 2020.

Gene Therapy and AAV Engineering

After my postbac year working on gene drives I joined a startup spinning out of the Church lab, Dyno Therapeutics. Along the way I helped assemble and characterize dozens of massive capsid libraries, some of which have become best in class for applications in the brain and eye.

Some of the work has also been presented at ASGCT. Below is just a selection of the projects.

• Efficient design of optimized AAV capsids using multi-property machine learning models trained across cells, organs and species. ASGCT.
• Optimizing Intravitreal Delivery to the Non-Human Primate Retina with Machine-Guided AAV Capsid Design. ASGCT.
• Machine-Guided Design Reveals AAV Variants Efficient in Transducing NHP Retina After Intravitreal Delivery. ASGCT.
• AAV Capsid Property Estimation Is Improved by Combining Single-Molecule ID Tags and Hierarchical Bayesian Modeling of Experimental Processes. ASGCT.
• Accurately Quantifying Transduction within Barcoded AAV Capsid Libraries via Tracking of Single-Molecule ID Tags. ASGCT.

Wearables

Continuous monitoring devices are necessary if we are to advance the current goals of AI in health and biology. The continuous glucose monitor (CGM) exists because of a natural enzyme we could ingeniously apply to an engineered system, but unfortunately nothing similar exists for other compounds. I worked at Caltech with Anand Muthuswamy under the advising of Henry Lester on applying frontier protein language and structure models towards the design of novel continuous monitors beyond the glucose monitor.

This work was part of a larger effort to create a new Focused Research Organization oriented on making new wearable devices.

Papers in process.