YuGene: a simple approach to scale gene expression data derived from different platforms for integrated analyses.

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Gene expression databases are crucial for understanding cell states.
• Assessing gene expression across different data platforms presents significant challenges.
• Existing data integration methods face barriers like format disparities, lack of common identifiers, and batch effects.

Purpose of the Study:

• To introduce YuGene, a novel cross-platform normalization strategy for gene expression data.
• To provide a simple transform that facilitates gene expression queries across diverse datasets.
• To address limitations of existing methods that compromise gene expression comparisons.

Main Methods:

• Developed the YuGene transform, assigning a modified cumulative proportion value to each measurement.
• Applied the transform to individual samples, accommodating varied data distributions and sizes.
• Benchmarked YuGene against common normalization methods like quantile (RMA), Z-score, and rank methods.

Main Results:

• YuGene performs favorably compared to established normalization approaches.
• The transform is robust to combining datasets of different sizes and does not require global renormalization.
• Implementation on www.stemformatics.org enables cross-dataset expression queries and provides probe performance statistics.

Conclusions:

• YuGene offers a flexible and effective solution for cross-platform gene expression analysis.
• The method simplifies data integration without losing essential distributional information.
• YuGene enhances the utility of gene expression databases for research, particularly in stem cell studies.