SNPrune: an efficient algorithm to prune large SNP array and sequence datasets based on high linkage disequilibrium.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• High pairwise linkage disequilibrium (LD) in single nucleotide polymorphism (SNP) data can impede genomic prediction model performance and efficiency.
• Pruning genotyping data for high LD is crucial for optimizing genomic prediction models.
• Existing methods for detecting SNPs in high LD can be computationally intensive.

Purpose of the Study:

• To develop an efficient algorithm, SNPrune, for rapid detection of SNPs in complete or high LD.
• To enable faster and more efficient pruning of large-scale genomic datasets.

Main Methods:

• Developed SNPrune algorithm to detect pairs of SNPs in complete or high LD.
• Algorithm sorts loci by minor allele count for efficient LD detection.
• Validated SNPrune against PLINK using SNP array and simulated whole-genome sequence data with an r² threshold of 0.99.

Main Results:

• SNPrune demonstrated comparable SNP removal to PLINK on pig data but was 12 times faster.
• On simulated sequence data, SNPrune identified millions of SNPs in complete and high LD.
• SNPrune was significantly faster than PLINK, especially with multi-threading, reducing analysis time considerably.

Conclusions:

• The SNPrune algorithm efficiently removes SNPs in high LD from large genomic datasets.
• SNPrune offers a computationally efficient solution for data pruning in genomics.
• This tool can accelerate the development and application of genomic prediction models.