Structure

Structure performs model-based Bayesian clustering to infer population structure from multi-locus genotype data, assign individuals to populations, estimate allele frequencies, and detect admixture and migrants while analyzing linkage disequilibrium patterns.

Key Features:

• Model-Based Clustering: Implements a probabilistic Bayesian framework that assumes an unknown number K of populations and estimates allele frequencies to assign individuals to one or more populations.
• Handling Linkage Disequilibrium: Accounts for linkage disequilibrium between loci to detect historical admixture, infer origins of chromosomal regions, and improve uncertainty estimates when loci are linked.
• Adaptability to Various Genetic Markers: Supports microsatellites and amplified fragment length polymorphisms (AFLPs) without assuming specific mutation processes.
• Handling Genotypic Ambiguity: Includes methods for dominant markers (e.g., AFLPs) to manage genotypic ambiguity, null alleles, and genotype-calling limitations in polyploids.
• Incorporation of Sample Group Information: Integrates sample group information by adjusting prior distributions for population assignments based on location-specific proportions.
• Cluster Number Estimation: Uses the log probability of the data alongside the ad hoc DeltaK statistic to assist estimation of the number of clusters K and detect hierarchical structure.

Scientific Applications:

• Population Structure Analysis: Infers and characterizes population structure from genotype data across species.
• Admixture Studies: Identifies admixed individuals and historical admixture events for studies of hybrid zones and migration.
• Genetic Diversity Surveys: Surveys genetic variation across multiple loci for conservation genetics and evolutionary analyses.
• Population Subdivision Detection: Detects subtle population subdivisions that may be missed by other methods.

Methodology:

Uses a probabilistic Bayesian clustering framework that assumes an unknown K and allele-frequency parameters, accounts for linkage disequilibrium, incorporates sample group priors, employs the DeltaK ad hoc statistic for cluster-number inference, and iteratively refines population assignments.