A comparative analytical assay of gene regulatory networks inferred using microarray and RNA-seq datasets

Area of Science:

• Genomics
• Bioinformatics
• Systems Biology

Background:

• Gene Regulatory Networks (GRNs) are crucial for understanding gene expression.
• High-throughput omics data (microarray, RNA-Seq) and preprocessing methods complicate GRN reconstruction.
• Guidelines are needed to assess the impact of data platforms and normalization on GRN accuracy.

Purpose of the Study:

• To compare the accuracy of GRNs reconstructed from microarray and RNA-Seq data.
• To evaluate the influence of different normalization procedures on GRN construction.
• To assess the topological similarities and differences between GRNs derived from distinct platforms.

Main Methods:

• Utilized public microarray and RNA-Seq datasets from Arabidopsis.
• Reconstructed six GRNs using RNA-Seq and microarray data with various normalization techniques.
• Performed comparative analysis and topological analysis of the reconstructed GRNs.

Main Results:

• Reconstructed GRNs were highly data-specific.
• RNA-Seq derived GRNs showed considerably higher accuracy than microarray-derived GRNs.
• Topological analysis revealed similarities but greater connectivity in RNA-Seq networks.
• Robust Multiarray Averaging (RMA) and Variance-Stabilizing Transformed (VST) normalization yielded the best performance.

Conclusions:

• RNA-Seq data provides more accurate GRN reconstruction compared to microarrays.
• Normalization methods significantly impact GRN accuracy, with RMA and VST being superior.
• Understanding platform and normalization effects is essential for reliable GRN inference.