Multiple Partial Regularized Nonnegative Matrix Factorization for Predicting Ontological Functions of lncRNAs.

Area of Science:

• Genomics
• Bioinformatics
• Molecular Biology

Background:

• Long non-coding RNAs (LncRNAs) are key regulators of biological processes and are implicated in complex diseases.
• Despite advances in sequencing, lncRNA functions remain largely unknown, hindering disease mechanism elucidation.
• Current prediction methods often rely on fusing heterogeneous genomic data, which may not fully capture lncRNA-function relationships.

Purpose of the Study:

• To develop a novel algorithm for predicting lncRNA functions.
• To overcome limitations of existing methods that fuse heterogeneous genomic data.
• To provide insights into the mechanisms of complex diseases by understanding lncRNA functions.

Main Methods:

• Proposed a nonnegative matrix factorization algorithm with multiple partial regularization (MPrNMF).
• Constructed lncRNA-gene associations separately for each genomic data type.
• Integrated these multiple associations using a regularization strategy instead of data fusion.

Main Results:

• The MPrNMF algorithm successfully predicted lncRNA functions without fusing heterogeneous genomic data.
• The proposed method demonstrated superior performance compared to state-of-the-art network-analysis methods.
• Validated the effectiveness of the model and algorithm in exploring lncRNA functions.

Conclusions:

• MPrNMF offers an effective approach for predicting lncRNA functions.
• The method provides a new avenue for investigating the roles of lncRNAs in complex diseases.
• This work enhances our understanding of lncRNA biology and its disease relevance.