Unveiling Commonalities and Differences in Genetic Regulations via Two-Way Fusion

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Understanding genetic regulation, including gene expressions (GEs) influenced by copy number variations and methylations, is vital for complex disease research.
• Recent studies explore diverse patient groups, moving beyond homogeneous cohorts to uncover disease heterogeneity.
• Analyzing complex regulatory networks where genes and regulators have multi-to-multi relationships presents significant computational challenges due to high dimensionality.

Purpose of the Study:

• To develop an integrative analysis approach for identifying commonalities and differences in genetic regulation patterns.
• To address the challenges posed by multi-gene regulation and multi-regulator effects in complex diseases.
• To provide a computationally efficient and robust method for analyzing high-dimensional genomic data.

Main Methods:

• A two-way fusion integrative analysis approach was developed, employing two fusion penalties to simultaneously analyze gene expression and regulator patterns.
• A Huber loss function was incorporated to enhance robustness against potential data contamination.
• An efficient iterative optimization algorithm was designed, ensuring global optimality without auxiliary variables or extra tuning parameters.

Main Results:

• The proposed approach effectively identifies commonalities and differences in both regulated gene expression patterns and regulator patterns.
• Extensive simulations demonstrated the advantages and effectiveness of the developed method.
• Application to The Cancer Genome Atlas data for melanoma and lung cancer yielded significant findings and strong predictive performance.

Conclusions:

• The novel two-way fusion method offers a powerful tool for dissecting complex genetic regulatory networks in diseases.
• This approach enhances our understanding of disease development and progression by revealing intricate gene-regulator relationships.
• The method shows promise for improving diagnostic and prognostic predictions in complex diseases like cancer.