Accurate continuous geographic assignment from low- to high-density SNP data

Area of Science:

• Genetics
• Statistics
• Bioinformatics

Background:

• Large-scale genotype data are crucial for tracking disease outbreaks and determining geographic origins.
• Applications span forensics, wildlife management, and epidemiology, but require efficient statistical tools.
• Existing methods struggle with the increasing volume of genetic information from modern sequencing technologies.

Purpose of the Study:

• To introduce a novel statistical method for geopositioning individuals based on their genotypes.
• To provide a fast and accurate tool for genetic ancestry inference.
• To assess the performance and scalability of the new method across different species and dataset sizes.

Main Methods:

• A geostatistical model trained on georeferenced genotype data was developed.
• Integrated Nested Laplace Approximation (INLA) was used for statistical inference, avoiding Monte Carlo simulations.
• The method's performance was compared to an alternative geospatial inference method (SPA) using simulated data.

Main Results:

• The novel method demonstrated accuracy in continuous spatial assignment across various scales.
• Analyses included genotype data from Florida Scrub-jay birds, Arabidopsis thaliana, and humans, with up to 197,146 single nucleotide polymorphisms (SNPs).
• The method is particularly well-suited for medium-sized datasets, such as reduced-representation sequencing data common in ecological studies.

Conclusions:

• The developed statistical method offers an accurate and efficient approach for genetic geopositioning.
• It addresses the need for advanced statistical tools to handle large genotype datasets.
• The method has broad applicability in population genetics, conservation, and forensic science.