Imaging-AMARETTO

Imaging-AMARETTO integrates multiomics, imaging, and clinical data to infer regulatory networks that link molecular drivers to imaging-derived biomarkers and clinical outcomes.

Key Features:

• Integration of Multi-Omics Data: Integrates genomic, transcriptomic, and other omics datasets with radiographic and histopathology imaging and clinical information.
• Network Biology Framework: Employs a graph-based fusion approach to link diseases, drivers, targets, and drugs.
• Regulatory Network Inference: Infers regulatory networks using integrative multi-omics analysis combined with penalized regression techniques.
• Imaging Biomarker Relevance: Correlates imaging-derived biomarkers from radiography and histopathology with molecular profiles to interpret mechanisms underlying clinical outcomes.
• Hypothesis Generation: Identifies known and novel master drivers and connects distinct biological pathways for downstream experimental validation.

Scientific Applications:

• Brain tumor imaging-genomics (GBM, LGG): Applied to glioblastoma multiforme (GBM) and low-grade glioma (LGG) by integrating multiomics from The Cancer Genome Atlas (TCGA) with radiographic imaging and Ivy Glioblastoma Atlas Project (IvyGAP) transcriptomics with histopathology imaging.
• Molecular driver discovery: Recapitulated microglia/macrophage-associated mechanisms mediated by STAT3, AHR, and CCR2, identified neurodevelopmental and stemness mechanisms involving OLIG2, and uncovered novel master drivers such as THBS1 and MAP2.

Methodology:

Implements a network-based imaging genomics approach that integrates multiomics data with imaging features, employs graph-based fusion and integrative multi-omics analysis, and uses penalized regression techniques to infer regulatory networks while integrating TCGA and IvyGAP datasets with radiographic and histopathology imaging.