A PCA-based method for ancestral informative markers selection in structured populations

Area of Science:

• Population Genetics
• Genomic Data Analysis
• Statistical Genetics

Background:

• Population structure analysis is crucial for understanding human history and identifying disease-related genes.
• Structured association (SA) methods are widely used but depend heavily on the quality and quantity of ancestral informative markers (AIMs).
• Existing AIM selection methods often require unavailable or uncertain prior individual ancestry information.

Purpose of the Study:

• To develop a novel, ancestry-information-free approach for selecting AIMs.
• To enhance the accuracy of population structure identification and association mapping.
• To provide a robust method for selecting informative AIMs from whole genome data.

Main Methods:

• Developed a new AIM selection method utilizing principle component analysis (PCA).
• The PCA-based approach does not require prior ancestry information from study subjects.
• Applied the method to both simulated and real genetic data.

Main Results:

• PCA-selected AIMs significantly improve the accuracy of inferred individual ancestries compared to random selection.
• The method demonstrates effectiveness with equivalent numbers of AIMs.
• Successfully applied to whole genome data for selecting highly informative AIMs.

Conclusions:

• The novel PCA-based AIM selection method overcomes limitations of existing approaches.
• This method enhances the accuracy of population structure inference and association studies.
• It offers a valuable tool for correcting population stratification biases in genetic analyses.