ResponderID™

A methodology for oncology RNA biomarker discovery at scale

Genialis biomarkers analyze gene expression data with an advanced AI foundation model and machine learning algorithms to predict treatment response and stratify patients. 

Over the past two decades of academic research and commercial development, Genialis scientists have honed a novel methodology to ensure its RNA-based biomarkers work across the entire translational and clinical lifecycle of drug development. We have repeatedly shown how biomarkers developed with our methodology perform on patient data from clinical trials and the real world. 

cancer patient
Genialis ResponderID is an advanced AI foundation model that maps the complex landscape of cancer biology - an AI recommendation engine to guide therapy for every cancer target, drug, and ultimately, patient.

At the heart of this approach is the

Genialis™ Supermodel

an advanced AI foundation model that maps the complex landscape of cancer biology. One might think of it like an AI recommendation engine to guide therapy for every cancer target, drug, and ultimately, patient.

  • The Supermodel runs on gene expression data, transforming the vast transcriptome into a subset of cancer biologies relevant to a drug target or mechanism.

  • A machine-learning algorithm is then trained on these biologies to yield a biology-driven predictive model.

  • Finally, this model is tuned and validated using patient data (clinical or real world), resulting in a highly accurate, deeply informative patient classifier that redefines the very word “biomarker.”

"We are so focused on getting the science right, because when we do, it changes lives."

Miha Štajdohar, PhD

CTO and Co-founder of Genialis

Key Advantages

Why is Genialis ResponderID better than other biomarker approaches?

  • We model biology, not clinical outcomes. This allows us to:

  • Develop, test, and refine biomarkers on preclinical and translational data, even before the first clinical trial results are collected, without sacrificing future accuracy in patient data
  • Integrate and extract value from relatively small datasets, e.g. those from clinical trials 
  • Create models that are generally useful across tissue histologies and accommodate multiple drug MOAs

  • We cover the entire lifecycle of a biomarker, including:

  • Precision medicine / clinical development strategy to maximize the value and impact of the biomarker program
  • Testing, tuning and validation to specific compound data to further differentiate the drug from its peers
  • Clinical trial assay/ Cdx development with our global network of diagnostic partners

  • This allows us to develop a new generation of biomarkers that are:

  • Accurate and interpretable, due to the information richness of RNA-seq
  • Truly representative, with globally diverse training data and a worlds-first RNA foundation model to fully capture complex biology
  • Scalable for the life of a drug, from cell lines to trial patients to the real world

First-in-class patient classifier

The first and only biomarker that can accurately stratify KRAS patients by clinical response

Use cases

ResponderID helps develop effective drugs and improve treatment decisions

Contact us to learn more about how ResponderID can build biomarkers for better treatment options