Comparison of Two Methods for Detecting Alternative Splice Variants Using GeneChip(®) Exon Arrays

Area of Science:

• Genomics
• Bioinformatics
• Molecular Biology

Background:

• Affymetrix GeneChip Exon Arrays are utilized for detecting alternative splice variants.
• Microarray Detection of Alternative Splicing (MIDAS) and Partek Genomics Suite (Partek GS) are popular analytical methods for exon array data.
• Differences in underlying assumptions and mathematical formulations between MIDAS and Partek GS complicate direct comparison.

Purpose of the Study:

• To compare the analytical methods Microarray Detection of Alternative Splicing (MIDAS) and Partek Genomics Suite (Partek GS) for analyzing exon array data.
• To investigate the impact of different assumptions on the detection of alternative splice variants.
• To evaluate the utility of a generalized probe model (GPM) in reconciling the results of MIDAS and Partek GS.

Main Methods:

• Analysis of a colon cancer exon array dataset using MIDAS, Partek GS, and a generalized probe model (GPM).
• Statistical significance testing was employed within MIDAS and Partek GS.
• The GPM was used to encompass both MIDAS and Partek GS under varying assumptions for comparative analysis.

Main Results:

• MIDAS and Partek GS produced distinct sets of genes identified with alternative splice variants.
• The GPM demonstrated results consistent with MIDAS and Partek GS individually, depending on the applied assumptions.
• Differences in results stem from MIDAS's focus on relative exon expression changes versus Partek GS's use of absolute exon expression values.

Conclusions:

• MIDAS and Partek GS offer complementary insights into alternative splicing.
• The choice of analytical method and its underlying assumptions significantly influence the discovery of alternative splice variants.
• Comprehensive alternative splicing analysis necessitates considering findings from both MIDAS and Partek GS.