AdmixTools

AdmixTools implements statistical methods to detect, quantify, and model population admixture using allele frequency and genotype data.

Key Features:

• Three-Population Test (f3-statistics): Uses f3-statistics to test whether a target population is admixed from two source populations by analyzing allele frequency data.
• D-Statistics: Calculates D-statistics to detect gene flow and introgression events that are not explained by simple tree-like evolutionary models.
• F4-Ratio Estimation: Estimates the ratio of genetic contributions from two ancestral populations to quantify ancestry proportions in an admixed population.
• Admixture Graph Fitting: Fits admixture graphs representing multiple events of gene flow and divergence to visualize and test hypotheses about historical population relationships.
• Rolloff Analysis: Analyzes the decay of linkage disequilibrium to estimate the number of generations since an admixture event.
• Bounds on Admixture Proportions (qpBound): Computes lower and upper bounds on admixture proportions using f3-statistics to constrain models of population mixture.
• Data Format Conversion (convertf): Converts data formats for compatibility with other programs in the suite.

Scientific Applications:

• Reconstruction of population history: Provides estimates of timing, proportions, and sources of genetic admixture to inform studies of evolutionary biology and human history.
• Large-scale SNP dataset analysis: Applied to a SNP array of 629,433 sites genotyped across 934 individuals from 53 populations to identify admixture signals, including in northern Europe involving populations related to present-day Basques and Sardinians and those related to northeast Asia and the Americas.
• Ancient DNA comparisons: Used alongside ancient DNA and archaeological data, including analyses of ancient bones from Sweden and the Tyrolean "Iceman," to investigate admixture between Neolithic migrants and indigenous Mesolithic populations in Europe.

Methodology:

Computational components include convertf for data format conversion; qp3Pop implementing the f3 three-population test; qpBound computing bounds on admixture proportions; qpDstat computing D-statistics; qpF4Ratio estimating F4-ratios; and rolloff analyzing linkage disequilibrium decay to estimate time since admixture.