fastSTRUCTURE

fastSTRUCTURE infers population structure from large-scale SNP genotype datasets using a variational Bayesian framework to provide rapid posterior estimates of ancestry proportions.

Key Features:

• Variational Bayesian Framework: Uses a variational Bayesian approach to approximate posterior distributions of ancestry proportions from SNP genotype data.
• Optimization-based Inference: Formulates posterior computation as an optimization problem to accelerate inference.
• Computational Efficiency and Speed: Achieves computational speeds nearly two orders of magnitude faster than STRUCTURE while maintaining accuracy comparable to ADMIXTURE.
• Heuristic Scores for K Selection: Incorporates heuristic scores to suggest a reasonable range for the number of populations (K) with minimal bias when structure is weak.
• Hierarchical Prior: Implements a hierarchical prior to enhance detection of subtle population structure.
• Validation on Simulated and Empirical Data: Validated on simulated datasets and real genotype data from the CEPH-Human Genome Diversity Panel.

Scientific Applications:

• Genetic Diversity Studies: Investigates genetic variation across different human populations or other species to assess genetic diversity.
• Evolutionary Biology: Assists reconstruction of historical migration patterns and evolutionary relationships between groups.
• Medical Genetics: Supports identification of population-specific genetic variants that may influence disease susceptibility.

Methodology:

Performs variational Bayesian posterior approximation by casting inference as an optimization problem and applies heuristic scores for K selection together with a hierarchical prior to improve detection of subtle structure.