A learning-based method to predict LncRNA-disease associations by combining CNN and ELM

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Long noncoding RNAs (lncRNAs) are crucial in biological processes and diseases.
• Traditional methods for identifying lncRNA-disease associations are time-consuming and costly.
• There is a need for efficient computational models to predict these associations.

Purpose of the Study:

• To propose a reliable and efficient computational model, LDACE, for predicting lncRNA-disease associations.
• To leverage deep learning technologies in bioinformatics for uncovering novel associations.
• To validate the model's performance and robustness.

Main Methods:

• Developed LDACE, a machine learning model combining Extreme Learning Machine (ELM) and Convolutional Neural Network (CNN).
• Constructed representation vectors by integrating functional and semantic similarity of lncRNAs.
• Utilized CNN for feature extraction and ELM for prediction of lncRNA-disease associations.

Main Results:

• Achieved high performance with an Area Under the ROC Curve (AUC) of 0.9086 (Leave-one-out) and 0.8994 (fivefold cross-validation).
• Case studies on lung and endometrial cancer demonstrated the model's robustness and efficiency.
• The model successfully predicted potential lncRNA-disease associations.

Conclusions:

• LDACE serves as a valuable auxiliary tool for guiding biomedical research.
• The integration of deep learning with biological big data offers novel insights into life sciences.
• The findings highlight the potential of computational approaches in understanding lncRNA-disease relationships.