Gene regulatory network inference based on modified adaptive lasso.

Area of Science:

• Bioinformatics
• Computational Biology
• Systems Biology

Background:

• Gene regulatory networks (GRNs) are crucial for understanding biological processes by illustrating gene interactions.
• Identifying direct gene relationships from high-dimensional, small-sample gene expression data presents a significant computational challenge.

Purpose of the Study:

• To develop an advanced GRN inference method that accurately identifies direct gene interactions.
• To enhance the precision of GRN construction by minimizing false positive edges and detecting both linear and nonlinear relationships.

Main Methods:

• Proposed a novel GRN inference method, Modified Adaptive Least Absolute Shrinkage and Selection Operator (MALasso).
• MALasso expands sample size using distance correlation and employs a new weighting strategy for adaptive lasso to refine network edges iteratively.
• Validated using simulated data and gene expression data from the DREAM challenge.

Main Results:

• MALasso demonstrated superior performance in Area Under the Receiver Operating Characteristic Curve (AUROCC) and Area Under the Precision-Recall Curve (AUPRC) compared to adaptive lasso and six other state-of-the-art methods.
• The method showed an improved capacity for distinguishing direct gene interactions from indirect ones.
• MALasso effectively detected linear and nonlinear relationships, reduced false positive edges, and increased accuracy in identifying direct gene relations.

Conclusions:

• The modified adaptive weighting in MALasso enhances the accuracy of GRN inference.
• MALasso offers a more precise tool for identifying direct gene regulatory relationships, advancing the field of systems biology.
• This method holds significant potential for dissecting complex gene interactions in various biological contexts.