Using genomic relationship likelihood for parentage assignment

Area of Science:

• Genomics
• Bioinformatics
• Quantitative Genetics

Background:

• Traditional parentage assignment relies on limited, independent genomic markers like microsatellites and low-density SNPs.
• Existing methods (exclusion-based, likelihood-based) assume marker independence, which is violated by dense SNP data.
• Genotyping errors and variable call rates complicate accurate parentage assignment with classical methods.

Purpose of the Study:

• To develop a fast and accurate trio parentage assignment method for dense SNP data.
• To create a method that does not require prior knowledge of genotyping error or call rates.
• To leverage genomic relationships for parentage inference, overcoming limitations of independent marker assumptions.

Main Methods:

• Developed the Genomic Relationship Likelihood (GRL) method, utilizing genomic relationships for parentage assignment.
• Applied GRL to simulated datasets with high-density SNPs, varying genotyping error and call rates.
• Compared GRL's accuracy and speed against established methods like Colony2 and an exclusion-based approach.

Main Results:

• GRL demonstrated high accuracy (approximately 99%) and speed in parentage assignment across simulated datasets.
• The method effectively handles dense, non-independent SNPs with variable call and genotyping error rates.
• GRL outperformed Colony2 in accuracy and provided a valuable tool for estimating genotype inconsistencies in exclusion-based models.

Conclusions:

• Genomic Relationship Likelihood (GRL) is a robust and efficient method for parentage assignment using dense SNP data.
• GRL overcomes the limitations of traditional methods by not requiring independent markers or prior error rate knowledge.
• The method offers a significant advancement for genetic research, breeding programs, and population genetics studies.