Data modeling as a main source of discrepancies in single and multiple marker association methods

Area of Science:

• Genetics
• Animal Breeding

Background:

• Genome-wide association studies (GWAS) are powerful for identifying disease loci in humans.
• High-density SNP maps in domestic animals enable accurate quantitative trait loci (QTL) detection via association studies.

Purpose of the Study:

• To analyze simulated data from the XII QTL-MAS meeting using association studies.
• To compare single marker association (SMA) and haplotype-based association (Blossoc) methods.
• To evaluate the impact of raw versus corrected data on QTL detection.

Main Methods:

• Applied SMA and Blossoc to raw and corrected datasets.
• Corrected data for infinitesimal, sex, and generation effects.
• Utilized bootstrap model aggregation to reconcile discrepancies.

Main Results:

• Both methods detected approximately ten significant loci, primarily on chromosomes 1, 4, and 5.
• Corrected data yielded more reliable results than raw data, reducing false positives.
• Bootstrap aggregation minimized discrepancies between raw and adjusted data analyses for SMA.

Conclusions:

• Careful consideration of model choice is crucial for robust GWAS.
• Data correction methods significantly influence the accuracy of QTL detection.