Dynamic model based algorithms for screening and genotyping over 100 K SNPs on oligonucleotide microarrays

Area of Science:

• Genomics
• Bioinformatics
• Population Genetics

Background:

• High-density single nucleotide polymorphism (SNP) coverage is crucial for population genetics and linkage studies.
• Millions of available SNPs and advancements in microarray technology present analytical challenges for genotyping.
• Developing robust methods for SNP selection and genotype calling is essential for large-scale genomic research.

Purpose of the Study:

• To introduce a novel dynamic model-based algorithm (DM) for efficient screening and genotyping of a large number of SNPs.
• To develop methods for selecting optimal SNP subsets and probes for microarray-based genotyping.
• To provide a high-quality genotype call with a reliable quality measure.

Main Methods:

• Developed a dynamic model-based algorithm (DM) with four underlying states (Null, A, AB, B) for probe quartets.
• Implemented SNP-level statistical aggregation across multiple probe quartets for genotype calling.
• Assessed algorithm performance using HapMap reference genotypes, Mendelian inheritance in families, and comparison with another genotype classification method.

Main Results:

• The DM algorithm screens over 3 million SNPs and genotypes over 100,000 SNPs.
• Achieved 99.81% concordance with HapMap reference genotypes at a 95.91% call rate across 1.5 million genotypes.
• Demonstrated a low Mendelian error rate of 0.018% in 10 trios and 99.90% consistency with MPAM.

Conclusions:

• The dynamic model-based algorithm provides a robust and accurate method for high-throughput SNP genotyping.
• The developed methods for SNP and probe selection enhance the efficiency of microarray-based genotyping.
• The DM algorithm is integrated into widely used genotyping software, facilitating its application in genomic research.