A Theoretical Approach for Correlating Proteins to Malignant Diseases

Area of Science:

• Biomedical Informatics
• Computational Biology
• Machine Learning in Medicine

Background:

• Malignant tumors develop over years due to complex biological factors, often initiated by habits and behaviors.
• The immune system may fail to recognize malignant tumors as foreign, complicating diagnosis.
• Identifying patterns between behaviors, environmental factors, and diseases is crucial for effective decision-making.

Purpose of the Study:

• To discover associations between normal proteins (environmental factors) and difficult-to-diagnose diseases.
• To propose justifications for the development of these diseases.
• To develop an efficient medical data mining technique for analyzing Erythrocytes Dynamic Antigens Store (EDAS) data.

Main Methods:

• Utilized machine learning techniques, specifically association rule mining.
• Proposed a modified Apriori algorithm for efficient data mining of EDAS data.
• Overcame limitations of existing algorithms like Apriori and Equivalence CLAss Transformation (ECLAT).

Main Results:

• Established a relationship between normal proteins (environmental, food, commensal, tissue) and disease proteins.
• Demonstrated associations between specific habits and behaviors with certain diseases.
• Validated the effectiveness of the modified Apriori algorithm for EDAS data analysis.

Conclusions:

• The study successfully identified significant associations between proteins, behaviors, and malignant diseases.
• The proposed data mining technique offers a more efficient approach for medical data analysis.
• The findings can assist clinical laboratories in uncovering biological causes of malignant diseases.