Network meta-analysis correlates with analysis of merged independent transcriptome expression data.

Area of Science:

• Bioinformatics
• Genomics
• Systems Biology

Background:

• High-dimensional transcriptome expression data from public repositories can be merged for novel group comparisons.
• Merging data from different technologies (microarray, RNA-seq) can introduce difficult-to-remove batch effects.
• Traditional meta-analysis faces challenges with biased results when merging diverse transcriptome datasets.

Purpose of the Study:

• To introduce and evaluate network meta-analysis for transcriptome expression data.
• To compare network meta-analysis with traditional data merging methods.
• To assess the utility of network meta-analysis in handling batch effects.

Main Methods:

• Network meta-analysis was applied to transcriptome expression data.
• A simulation study was conducted to compare network meta-analysis with merged data analysis.
• The method was validated using a real-world neuroinfection research dataset.

Main Results:

• Network meta-analysis results showed high correlation with merged data analysis in simple networks.
• Network meta-analysis yielded fold changes closer to simulated values when supported by multiple studies.
• The practicability of network meta-analysis was demonstrated on a real-world dataset.

Conclusions:

• Network meta-analysis is a valuable tool for inferring new insights from combined molecular expression studies.
• This method is particularly beneficial for overcoming batch effects in high-throughput expression data.
• Network meta-analysis provides a robust alternative for analyzing complex, multi-study transcriptome datasets.