GAPIT version 4: integration of GWAS into genomic prediction

Area of Science:

• Quantitative genetics
• Genomic prediction
• Statistical genomics

Background:

• Genomic prediction uses all markers, regardless of GWAS significance.
• Advancements in GWAS methods suggest integrating results can enhance prediction accuracy.

Purpose of the Study:

• To evaluate the effectiveness of incorporating GWAS findings into genomic prediction.
• To compare different GWAS models for their impact on prediction accuracy.

Main Methods:

• The study utilized the Genomic Association and Prediction Tool (GAPIT) to implement GWAS-Assisted Genomic Best Linear Unbiased Prediction (GAGBLUP).
• Simulations and real trait data were used to compare GAGBLUP with traditional genomic Best Linear Unbiased Prediction (GBLUP).
• Multiple-locus GWAS models (e.g., BLINK) were compared against single-locus models.

Main Results:

• GAGBLUP's effectiveness is dependent on the GWAS model used.
• Multiple-locus models, particularly BLINK, outperformed single-locus models.
• Incorporating BLINK GWAS results into GBLUP improved prediction accuracy by over 20% compared to GBLUP alone.

Conclusions:

• GAGBLUP effectively integrates GWAS insights with GBLUP's polygenic modeling.
• This approach offers more stable predictions across diverse genetic backgrounds.
• The enhanced utility of genomic selection is demonstrated for broader applications in breeding programs.