Cyclomics
Epinn®
Next generation Tumor classification
Rapid tumour classification to assist the pathologist during diagnosis, shortening the timeline from 1 week to <90 minutes! Which enables rapid decision-making for the clinicians, and makes it possible to classify tumours during surgery. Epinn enables AI-powered methylation analyses across ONT, Illumina, and bisulfite sequencing workflows.
Neural network
A compact three-layer classifier trained entirely offline on 36.8 million simulated runs. All the heavy computation happens before classifying.
Adaptive sampling
A nanopore-specific technique that reads the first bases of a DNA strand, checks if it maps to a region of interest, and ejects it early if not. This boosts the yield of informative CpG sites
Platform agostic
Epinn performs consistently across different sequencing hardware, flowcell chemistries, so it isn't tied to one specific sequencer generation.
Patient agnostic
Unlike earlier approaches that required training a new model on each patient's own data during surgery, Sturgeon uses one pretrained, universal model that works out of the box on any new sample.
Epinn® in a nutshell
Epinn is Cyclomics' framework for deploying AI-driven tumor classification in the clinic. It's built around neural networks trained to classify tumors from the sparse, unpredictable methylation data. The models of Epinn get around this by training on millions of simulated sparse sequencing runs, so it learns to recognize tumor signatures even from partial data. This has been clinically validated and is already used in surgery to guide decisions in real time.
​
Epinn's role is to make that power usable: it wraps tumour classification models in a simple graphical interface, so clinicians and researchers can run classification without coding or ML expertise, taking a sample to result in about 90 minutes.
Resources
Oncode Institute
Cyclomics secures an exclusive worldwide license from Oncode to Sturgeon software, technology, and IP, enhancing its Epinn platform for rapid DNA methylation–based cancer classification
Oncode Institute
Oncode Investigator Jeroen de Ridder & Cyclomics receive Health~Holland PPP subsidy to continue developing brain tumor classifier
Nat Geo
Meet Sturgeon, the AI tool that helps doctors identify brain tumors faster than ever
Nature
Ultra-fast deep-learned CNS tumour classification during surgery