Analyzing a co-occurrence gene-interaction network to identify disease-gene association

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Understanding genetic networks is crucial for chronic disease research, particularly cancer.
• Gene-gene interactions play a significant role in disease pathogenesis.
• Identifying these interactions aids in understanding complex diseases.

Purpose of the Study:

• To develop a text mining system for constructing a human genome-wide gene-gene interaction network.
• To identify disease-related genes, specifically for cancer, using network analysis.
• To evaluate the system's accuracy in predicting gene-gene interactions and disease associations.

Main Methods:

• Utilized text mining to recognize interacting genes based on co-occurrence in biomedical literature.
• Employed linear and non-linear rare-event classification models for gene interaction recognition.
• Applied network centrality measures (betweenness, closeness, eigenvector, degree) to rank gene importance and identify potential disease genes.

Main Results:

• The system achieved high average precisions (80-100%) in identifying genes associated with breast, prostate, and lung cancer.
• A prostate cancer case study demonstrated the system's ability to predict an average of 80% of prostate-related genes.
• The network analysis successfully ranked central genes, highlighting potential cancer-related genes.

Conclusions:

• The developed text mining system shows potential for enhancing the accuracy of gene-gene interaction and disease-gene association predictions.
• The system's network analysis approach effectively identifies key genes involved in disease.
• Further validation and comparison with state-of-the-art methods confirm the system's utility.