Genialis krasID

Genialis™ krasID is an AI/ML-powered biomarker platform designed to predict and monitor drug response across the full spectrum of research contexts — from preclinical models to intent-to-treat real-world patient cohorts. Built specifically around KRAS-associated biology, krasID incorporates over 20 AI/ML biomodules to generate a fundamentally new type of biomarker that reflects underlying biological states rather than relying solely on genetic variants such as mutation status.

krasID is intended for oncology drug developers, translational researchers, and clinical teams working with KRAS inhibitors. The platform is customizable for different KRAS inhibitors, tissue types, and treatment combinations, enabling it to be tuned to specific preclinical or clinical contexts. This flexibility allows the resulting biomarkers to differentiate a compound against a competitive field of KRAS-targeted therapies.

Core Capabilities

• Identify patients most likely to benefit from a given KRAS inhibitor and provide biological rationale for predicted response
• Predict the duration of response to KRAS inhibitor therapy
• Elucidate early disease progression in KRAS-mutated cancers treated with KRAS inhibitors
• Generate RNA-based phenomarkers that capture biological state rather than genetic variant alone
• Support monotherapy or combination therapy strategy decisions
• Identify clinically actionable resistance mechanisms

Validated Performance

• Real-world data validation: Patient data from real-world sotorasib-treated cohorts mirror outcomes observed in the CodeBreaK100/200 clinical trials, confirming the platform's translational relevance
• Extended time on treatment: Patients identified as responders by krasID remained on sotorasib nearly 50% longer than the broader cohort of KRAS G12C-selected patients
• High predictive accuracy: krasID predicts response with greater than 94% accuracy in preclinical models and greater than 80% accuracy in real-world patient populations

Scientific Foundation and Customization

• krasID leverages the Genialis™ Supermodel, a foundation model of cancer biology used to derive novel biomarkers for diagnosis, clinical development, and treatment decision-making in oncology
• RNA phenomarkers are central to the platform's approach, capturing dynamic biological states that genetic variant testing alone cannot reveal
• The platform can be tuned for a particular compound, tissue type, or clinical question, making it adaptable across drug development programs
• Algorithms have been validated on real-world results and presented at major oncology conferences including ASCO 2025, AACR 2025, and the RAS-Targeted Drug Development Summit 2024

Genialis krasID is developed by Genialis, Inc., with offices in Houston, Ljubljana, and Boston. The company is a founding member of the Alliance for AI in Healthcare. Supporting scientific documentation, including a detailed white paper and conference posters from ASCO 2025, AACR 2025, and RAS Summit 2024, is available directly from Genialis.