Evaluation of artificial time series microarray data for dynamic gene regulatory network inference.

Area of Science:

• Systems Biology
• Bioinformatics
• Computational Biology

Background:

• High-throughput technologies, such as microarrays, are essential for inferring gene regulatory networks (GRNs).
• Analyzing time-series data is crucial for understanding GRN dynamics and identifying dynamic networks.

Purpose of the Study:

• To evaluate the information content in artificial time-series microarray data for GRN inference.
• To assess the accuracy of models produced by inference processes using this data.
• To identify the limitations imposed by microarray data characteristics on GRN inference.

Main Methods:

• Dynamic artificial gene regulatory networks were used to generate artificial microarray data.
• Key data features (time separation, percentage of triggered genes, triggering function type) were systematically altered.
• Factors influencing inference performance (network size, noise, sparseness) were examined using a system theory approach.

Main Results:

• Time separation and the percentage of directly triggered genes were identified as crucial factors for accurate GRN inference.
• Network sparseness, triggering function type, and noise in input data significantly affect inference performance.
• Interactions between parameters were observed, where altering one factor impacted the dynamic response of others.

Conclusions:

• The number of datasets is the most significant parameter for GRN inference performance.
• Simulation findings were validated using a real GRN, confirming the importance of identified factors.
• Understanding data characteristics and parameter interactions is vital for robust dynamic GRN inference.