dXY

dXY is a population genetics statistic that measures the average number of nucleotide differences per site between two populations. It quantifies how genetically divergent the populations are by considering sequence data from individuals drawn from each population.

Calculation and interpretation: For two populations X and Y with nX and nY sequences, examine each site across aligned sequences. At site i, compute the proportion of all X–Y pairs that differ, denoted dXY(i). This is the number of XY pairs with different nucleotides at site i divided by nX multiplied by nY. The overall dXY is the average of dXY(i) across all sites (often across sites with data). Higher dXY values indicate greater sequence divergence between the populations, while lower values indicate closer relatedness or recent gene flow.

Relation to other diversity statistics: dXY is distinct from within-population measures. Let dx be the average pairwise difference within population X, and dy within Y. A commonly used related quantity is the net divergence da, defined as dXY minus the average within-population diversity (dx + dy)/2. This distinction helps separate historical divergence from current variation within populations.

Applications: dXY is used in comparative genomics and population-genetics studies to assess population structure, infer split times, and identify regions of the genome with elevated or reduced divergence. Genome-wide scans can reveal loci under selection, barriers to gene flow, or introgression events by examining local dXY values across sliding windows.

Limitations: dXY can be influenced by sample size, missing data, and uneven sequencing depth. It is also affected by differences in mutation rate and recombination, and it reflects both historical divergence and contemporary processes such as gene flow. Proper interpretation often requires complementary statistics like FST, dx, dy, and da, as well as demographic context.