From PoC to Platform: The Journey so far

We started with a feasibility study. With the help of cancer genomics expert Dr. Elena Tomas Bort, we surveyed public bisulfite sequencing datasets across various cancers. Hepatocellular carcinoma (HCC) emerged as a compelling proof-of-concept—not only because of personal urgency, but because it lacks good screening standards today. Our team—including bioinformatics consultant Paul Gaudin—built a preprocessing pipeline to clean, align, and format raw sequencing data from multiple independent studies. The challenge? Each dataset used a different assay and targeted different methylation markers. There was minimal overlap.

That’s when the insight struck: instead of one universal model, we would build multiple high-accuracy models, each tailored to its dataset, and use a meta-classifier to route samples based on assay signature. This architecture—validated using an independent fourth dataset—delivered strong performance, confirming that our platform could generalize across assays and patients.

Dr. Ryan Koehler, a computational biologist advisor, helped ensure scientific rigor and reproducibility of our approach. Together, we built the foundation for OmicXScreen, a modular platform designed to scale to multi-cancer detection.

What's Next

We’re now gearing up for clinical validation of OmicXScreen, with plans to launch as a CLIA-certified Laboratory Developed Test (LDT) in early 2026—starting with liver cancer and cirrhosis.

But this is just the beginning. Our modular AI/ML platform will enable personalized, multi-cancer screening from a single blood draw, tailored to individual risk factors and clinical context. We believe this can move cancer detection upstream—where it belongs—into preventive care and primary care workflows.

Thank you for following our journey. More to come.


Mythili Subharam
Founder/CEO

June 2025