PSnoD

PSnoD predicts associations between small nucleolar RNAs (snoRNAs) and human diseases using matrix completion with bounded nuclear norm regularization to prioritize candidate snoRNA–disease links for biomedical research.

Key Features:

• Bounded Nuclear Norm Regularization: Implements bounded nuclear norm regularization within the matrix completion framework to improve prediction accuracy.
• Matrix Completion-based Prediction: Infers missing snoRNA–disease associations from existing association data using matrix completion methods.
• Benchmark Evaluation: Evaluated with 5-fold stratified shuffle split, achieving an area under the ROC curve of 0.90 and an area under the precision-recall curve of 0.55.
• Computational Efficiency: Reported to outperform other matrix completion techniques in computational speed and resource usage.

Scientific Applications:

• Identification of snoRNA–disease associations: Enables prioritization of candidate snoRNA–disease links for experimental validation.
• Understanding disease mechanisms: Supports investigation of snoRNA roles in disease pathogenesis.
• Biomarker discovery: Aids identification of novel biomarkers for disease diagnosis and prognosis.
• Therapeutic target discovery: Guides discovery of therapeutic targets related to snoRNA biology.

Methodology:

Applies matrix completion using bounded nuclear norm regularization to predict associations from existing snoRNA–disease data, with performance assessed by 5-fold stratified shuffle split (ROC AUC 0.90, PR AUC 0.55).