Spatial autocorrelation analysis of individual multiallele and multilocus genetic structure

Area of Science:

• Population genetics
• Spatial autocorrelation analysis
• Molecular ecology

Background:

• Population genetic theory predicts spatial autocorrelation due to restricted gene flow.
• Empirical studies often show weak or inconsistent spatial structure across loci and sites.
• Existing methods analyzing individual alleles may lack statistical sensitivity for complex genetic data.

Purpose of the Study:

• Introduce a novel multivariate approach for spatial autocorrelation analysis.
• Develop a method applicable to multiallelic, codominant, multilocus genetic data.
• Enhance the detection of spatial genetic structure by treating the entire dataset holistically.

Main Methods:

• Developed a general multivariate method based on genetic distance calculations.
• Applied the method to multiallelic codominant loci.
• Incorporated nonparametric permutational testing for correlogram analysis.
• Utilized an example dataset from the orchid Caladenia tentaculata.

Main Results:

• The multivariate approach strengthens spatial signals and reduces noise compared to single-allele analyses.
• Highly polymorphic loci with intermediate frequency alleles yielded clearer results than rare alleles.
• Multilocus analysis provided more robust conclusions than single-locus treatments.
• Differential allele weighting offered minimal improvement in resolution.

Conclusions:

• The new multivariate method effectively detects spatial genetic structure in plant populations.
• This approach offers improved resolution and reliability for analyzing complex genetic marker data.
• Findings encourage broader application of this method in population genetics and conservation.