Prioritizing Susceptible Genes for Thyroid Cancer Based on Gene Interaction Network

Area of Science:

• Endocrinology
• Oncology
• Bioinformatics

Background:

• Thyroid cancer is the second most common endocrine malignancy.
• Early-stage thyroid cancer is often asymptomatic, delaying diagnosis and treatment.
• Genetic factors play a crucial role in thyroid cancer development.

Purpose of the Study:

• To identify core genes driving thyroid cancer through network analysis.
• To develop a more accurate method for detecting thyroid cancer-related genes.
• To improve early diagnosis and treatment strategies for thyroid cancer.

Main Methods:

• Construction of a thyroid cancer-related gene interaction network.
• Application of random walks for network traversal.
• Utilizing ADNN (Adaboost and deep neural network) for gene sorting and discovery.
• Validation through fivefold cross-validation.

Main Results:

• Identification of key interacting genes within the thyroid cancer network.
• ADNN demonstrated superior performance in gene identification compared to other methods.
• ADNN achieved an Area Under the Curve (AUC) of 0.85 and an Area Under the Precision-Recall Curve (AUPR) of 0.81.
• Discovery of novel thyroid cancer-related genes using the ADNN model.

Conclusions:

• The interaction network approach effectively highlights core genes in thyroid cancer.
• ADNN is a highly accurate and effective tool for identifying and discovering thyroid cancer-related genes.
• This research provides a foundation for improved diagnostic strategies combining genetic and imaging data.