[back to the DMCM’25 Programme]

Title:
Knowledge-based methods to extract disease mechanisms from multi-omics data

Authors:
Denes Turei and Saez-Rodriguez Group

Abstract:
In the Saez-Rodriguez Group we develop methods to extract mechanistic insights from omics data and prior knowledge. In our typical workflows we infer activities of molecular actors, such as transcription factors or pathways, based on their signatures available in databases. The activity inference largely reduces dimensionality, and improves robustness and interpretability. Next, we map the activities to causal, context agnostic prior knowledge networks, and apply network optimization methods to infer context specific mechanisms, i.e. networks that best explain the activities under each specific condition, such as in disease or drug effect. Our approach also offers a flexible way to integrate multiple omics modalities, as activities derived from different omics can be analysed together once mapped on a unified causal network, covering both signaling and metabolism. Over the past years, we not only our made solutions for prior knowledge access, activity inference and network optimization more user-friendly and reliable, but also improved their integration between each other, and with frameworks for sincle-cell, cell-cell communication, spatially resolved and multi-omics analysis. Here we present an overview of our tools, their use cases and future developments.