NSCGRN: a network structure control method for gene regulatory network inference.

Area of Science:

• Computational Biology
• Systems Biology
• Bioinformatics

Background:

• Gene regulatory network (GRN) inference is crucial for understanding disease mechanisms.
• Identifying redundant regulatory elements remains a significant challenge in computational GRN analysis.
• Existing methods combining global and local network topologies have limitations in defining their specific forms and cooperation.

Purpose of the Study:

• To propose a novel network structure control method (NSCGRN) for gene regulatory network inference.
• To define specific forms and a cooperative mode for global and local network topologies in GRN inference.
• To improve the accuracy and reduce redundancy in GRN inference.

Main Methods:

• Developed the network-structure-controlling-based GRN inference method (NSCGRN).
• Employed a 'global topology dominates and local topology refines' cooperative mode.
• Utilized network topology centrality sorting, Bernaola-Galvan mutation detection, and integration of four network motifs (fan-in, fan-out, cascade, feedforward loop).

Main Results:

• NSCGRN achieved the highest F1 and Matthews correlation coefficient scores compared to state-of-the-art methods across three datasets (six networks).
• Demonstrated unique advantages in gene regulatory network inference.
• Successfully controlled upstream and downstream regulations within a global scope and optimized local complex regulations.

Conclusions:

• The proposed NSCGRN method offers a robust approach to gene regulatory network inference.
• The specific forms and cooperative mode of global and local topologies enhance GRN analysis.
• NSCGRN provides a significant advancement for understanding gene regulation and disease pathogenesis.