Comparison and development of cross-study normalization methods for inter-species transcriptional analysis

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Joint analysis of gene expression datasets across experiments is challenging due to technical variations.
• Existing cross-study normalization methods are not designed for inter-species data analysis.
• Inter-species gene expression analysis requires methods that account for both experimental and biological variability.

Purpose of the Study:

• To evaluate the performance of existing cross-study normalization methods (EB, DWD, XPN) for inter-species RNA sequencing data.
• To develop a novel normalization method specifically for cross-study and cross-species gene expression analysis.
• To create an evaluation framework for assessing normalization methods' ability to reduce experimental effects while preserving biological signals.

Main Methods:

• Applied three established cross-study normalization methods (EB, DWD, XPN) to inter-species RNA sequencing datasets.
• Developed a novel performance evaluation approach for cross-study normalization.
• Proposed and implemented a new cross-study and cross-species normalization method (CSN).

Main Results:

• Existing methods (EB, DWD, XPN) showed moderate success in cross-species normalization.
• XPN excelled at reducing experimental differences, while EB was better at preserving biological signals.
• The newly developed CSN method demonstrated superior performance in balancing the reduction of experimental effects and the preservation of biological differences.

Conclusions:

• Cross-study normalization techniques are applicable and beneficial for inter-species gene expression analysis.
• The proposed CSN method offers an improved approach for normalizing inter-species datasets, enhancing biological data integrity.
• This work paves the way for developing more advanced normalization strategies for comparative multi-species genomic studies.