Unbiased microRNA-Disease Association Prediction Using ICD-11 Codes and Negative Sampling

Area of Science:

• Bioinformatics
• Genomics
• Computational Biology

Background:

• MicroRNA-disease associations are crucial for understanding disease mechanisms and developing targeted therapies.
• Existing computational models often rely on similarity-based methods, limiting their applicability to novel or understudied microRNAs and diseases.
• The lack of comprehensive negative sample datasets hinders accurate prediction model development.

Purpose of the Study:

• To develop a novel computational model, Unbiased microRNA-disease association predictor (UBMDA), for predicting microRNA-disease associations.
• To overcome limitations of previous methods by utilizing International Classification of Diseases (ICD-11) codes and microRNA nucleotide sequences as input features.
• To construct a balanced negative sample dataset that accounts for potential biases in microRNA and disease frequencies.

Main Methods:

• Developed UBMDA, a computational model employing ICD-11 disease codes and microRNA nucleotide sequences for feature extraction.
• Created a negative sample dataset by carefully considering the frequencies of microRNAs and diseases present in the positive sample dataset to prevent prediction bias.
• Implemented a strategy to ensure similar microRNA and disease frequencies between positive and negative sample datasets.

Main Results:

• Successfully developed the UBMDA computational model with a simple and intuitive structure.
• Demonstrated the model's capability to predict microRNA-disease associations without relying on similarity-based feature extraction.
• The approach addresses the challenge of limited negative sample data in microRNA research.

Conclusions:

• UBMDA offers a novel approach for predicting microRNA-disease associations, applicable to newly discovered or poorly characterized microRNAs and diseases.
• The model's design, utilizing ICD-11 codes and nucleotide sequences, enhances its versatility and predictive power.
• UBMDA is expected to accelerate the discovery of microRNA-related biomarkers and therapeutic strategies.