The power of single-nucleotide polymorphisms for large-scale parentage inference

Area of Science:

• Genetics
• Bioinformatics
• Computational Biology

Background:

• Parentage inference is crucial for population genetics and conservation.
• Likelihood-based methods are powerful but computationally intensive.
• Approximating likelihood-ratio statistics is key for large-scale applications.

Purpose of the Study:

• To develop efficient importance-sampling algorithms for likelihood-based parentage inference.
• To estimate false-positive rates and assess the power of single-nucleotide polymorphisms (SNPs) for parentage studies.
• To compare likelihood-based methods with exclusion-based methods.

Main Methods:

• Developed and applied importance-sampling algorithms to approximate tail probabilities of the likelihood-ratio statistic.
• Utilized single-nucleotide polymorphism (SNP) data for parentage inference simulations.
• Investigated the impact of genotyping errors and related individuals on inference accuracy.

Main Results:

• Importance-sampling algorithms can accelerate tail probability computations over a millionfold.
• 60-100 SNPs are sufficient for accurate pedigree reconstruction in large populations.
• Likelihood-based inference is significantly more powerful than exclusion-based methods, requiring fewer SNPs.

Conclusions:

• Importance-sampling algorithms enhance the efficiency of likelihood-based parentage inference.
• SNPs are highly effective for large-scale parentage studies in managed and natural populations.
• Likelihood-based methods offer superior accuracy and efficiency compared to exclusion-based approaches.