Case Study · DiFeo Lab Benchmark

Tested by an independent lab.
The results outperformed the commercial benchmark.

Novarna partnered with the University of Michigan’s DiFeo Lab to design RNA candidates against a live ovarian cancer target. The DiFeo Lab ran the experiments independently. The designs outperformed a commercially available linear siRNA.

See the results How we worked together

Two DiFeo Lab researchers running experiments at the bench in the University of Michigan Department of Pathology. — Photo: U-M Department of Pathology · DiFeo Lab

Key takeaways

Outperformed

The commercially available linear siRNA for the same cancer target.

Wet lab

Results confirmed in cancer cell models by an independent research team.

Patent

Eligible molecule types generated directly from the AI design process.

The stakes

Why this research matters.

High-grade serous ovarian cancer carries a five-year survival rate of approximately 28%. Treatments that more precisely target the signals driving that disease have the potential to change that number. Novarna doesn’t run clinical trials.

But every candidate we design is a step toward a treatment that might.

The challenge

Ovarian cancer cells use a specific signal to grow. The DiFeo Lab needed a better way to block it.

The DiFeo Lab studies how ovarian cancer cells survive and spread. A key target is a molecular signal cancer cells exploit to shut down the body’s natural defenses — allowing tumors to grow.

Existing treatments were insufficient — too weak, too inconsistent. Traditional design methods had no answer. The DiFeo Lab had been looking for a better one. So had we.

How we worked together

Novarna designed the candidates. The DiFeo Lab tested them.

Novarna and the DiFeo Lab established a common target

Novarna and the DiFeo Lab aligned on the target — a specific molecular signal in ovarian cancer cells — and on what a successful design would need to do.

NovaEngine™ generated RNA candidates from scratch

Novarna’s AI generated RNA molecules specifically for this target — not adapted from existing treatments, designed from scratch against the target’s biological profile.

NovaLab™ synthesized and prepared the candidates

Novarna’s wet lab synthesized the molecules and handed them off to the DiFeo Lab ready for testing.

DiFeo Lab ran independent testing in cancer cell models

The DiFeo Lab tested the candidates in their ovarian cancer cell models — the same experimental conditions used to evaluate any other candidate treatment.

Results confirmed — exceeded the benchmark

The designs outperformed a commercially available linear siRNA — and produced new molecule types with patent potential.

The results

Novarna’s designs outperformed the commercial benchmark — in an independent lab.

Tested in cancer cell models — the same standard used to evaluate real treatments.

All testing conducted by the University of Michigan team — not by Novarna.

Produced new molecule types with patent potential.

On-target percentage efficiency

Representative data

~4%

Negative
control

~48%

Commercial
linear siRNA

~82%

Design A
Novarna

~93%

Design B
Novarna

The DiFeo Lab measured how well each candidate blocked disease activity — comparing Novarna’s molecules against the commercially available linear siRNA.

Your target. Our engine.

Want to see what we’d design for your target?

Share your target with our science team — we’ll scope a program against your specific biology.