Abstract P30: Q-omics: smart data mining through ontology-derived gene network modelling

Abstract

Q-omics (https://qomics.io) is an AI-powered omics data mining platform designed to help cancer researchers and oncologists identify tumor-specific targets, prognostic biomarkers, and underlying mechanisms—without requiring bioinformatics expertise. The platform integrates over 1 billion multi-modal omics data points to compute more than 30 billion high-confidence associations, reinforced by pan-cancer consensus scores across diverse methods, sample types and experiments. These associations are stored in MetaDBs, forming a foundational resource for robust biomarker discovery and machine learning applications. At the core of Q-omics is an ontology-derived gene network model that detect functional modules and gene hotspots based on quantitatively processed ontologies (GO, HPO and gene sets from MSigDB). This enables prediction of clinically relevant targets and mechanistic insights across cancer types. Leveraging large language models, Q-omics allows users to perform creative "text-to-data mining" transforming natural-language queries into automated workflows, analytical results and interpretation. Q-omics is now evolving toward a foundational Large Data Model (Bio-LDM) that supports autonomous data interpretation and novel hypothesis generation across user-submitted omics datasets.