A multi-layer inference approach to reconstruct condition-specific genes and their regulation.

Area of Science:

• Systems Biology
• Genomics
• Bioinformatics

Background:

• Reconstructing context-dependent gene networks is crucial for understanding systems biology.
• Identifying condition-specific regulatory mechanisms presents a significant challenge due to complex gene interactions.

Purpose of the Study:

• To develop and validate a multi-layer approach for reconstructing condition-specific gene regulatory networks.
• To identify novel target genes and transcription factor regulations in human breast cancer and plant cold acclimation.

Main Methods:

• Integrative analysis of gene expression, protein interaction, and transcription factor-target gene data.
• Development of a multi-layer computational framework for network reconstruction.
• Validation using synthetic datasets, yeast data, and experimental confirmation.

Main Results:

• Accurate reconstruction of condition-specific gene networks demonstrated on synthetic and yeast datasets.
• Identification of TACSTD2 (TROP2) as a target gene in human breast cancer, regulated by CREB and NFkB.
• Prediction of KIF2C as a target gene in ER-/HER2- breast cancer, regulated by E2F1, with experimental validation.

Conclusions:

• The multi-layer approach effectively reconstructs condition-specific gene networks.
• The study provides novel insights into gene regulation in human breast cancer and plant stress responses.
• Validated predictions highlight the utility of the approach for biological discovery.