The VKORC1 Asp36Tyr variant and VKORC1 haplotype diversity in Ashkenazi and Ethiopian populations

Abstract

The vitamin K epoxide reductase (VKORC1) is a key enzyme in the vitamin K cycle impacting various biological processes. VKORC1 genetic variability has been extensively studied in the context of warfarin pharmacogenetics revealing different distributions of VKORC1 haplotypes in various populations. We previously identified the VKORC1 Asp36Tyr mutation that was associated with warfarin resistance and with distinctive ethnic distribution. In this study, we performed haplotype analysis using Asp36Tyr and seven other VKORC1 markers in Ashkenazi and Ethiopian-Jewish and non-Jewish individuals. The VKORC1 variability was represented by nine haplotypes (V1-V9) that could be grouped into two distinct clusters (V1-V3 and V4-V9) with intra-cluster difference limited to two nucleotide changes. Phylogeny analysis suggested that these haplotypes could have developed from an ancestral variant, the common V8 haplotype (40 % in all population samples), after ten single mutation events. Asp36Tyr was exclusive to the V5 haplotype of the second cluster. Two haplotypes V5 and V4, distinguished only by Asp36Tyr, were prevalent in both Ethiopian population samples. The V2 haplotype, belonging to the first cluster, was the second most prevalent haplotype in the Ashkenazi population sample (15.8 %) but relatively uncommon in the Ethiopian origin (4.5-4.7 %). We discuss the genetic diversity among studied populations and its potential impact on warfarin-dose management in certain populations of African and European origin.

Appendix: Mutation frequency ratio and population admixture

Let A and B be two populations, m _A and m _B the admixture portions (per generation) in each of the populations, respectively. Let p _A(t) and p _B(t) be population frequencies of mutation (a SNP allele) in generation t. Then these two frequencies will be gradually equalized from generation to generation according to the following formulae.

$$ \begin{array}{l}{p}_A\left(t+1\right)={p}_A(t)\ast \left(1-{m}_A\right)+{p}_B(t)\ast {m}_A\hfill \\ {}{p}_B\left(t+1\right)={p}_B(t)\ast \left(1-{m}_B\right)+{p}_A(t)\ast {m}_B\hfill \end{array} $$

(A1)

As a simple consequence that can be deduced from (A1) by induction, it follows that after t generations of admixture the frequencies equal:

$$ \begin{array}{l}{p}_A(t)={p}_A(0)\ast \left(1-{m}_A\ast \Delta (t)\right)+{p}_B(0)\ast {m}_A\ast \Delta (t)\hfill \\ {}{p}_B(t)={p}_B(0)\ast \left(1-{m}_B\ast \Delta (t)\right)+{p}_A(0)\ast {m}_B\ast \Delta (t)\hfill \end{array}, $$

(A2)

where $\Delta (t)=\frac{1-{\left(1-{m}_A-{m}_B\right)}^t}{m_A+{m}_B}$, and p _A (0) and p _B (0) are the initial mutation frequencies in two populations.

When the mutation is absent in population A in the initial generation, p _A (0) = 0, the ratio between frequencies is not dependent on the initial mutation frequency in population B:

$$ \raisebox{1ex}{${p}_A(t)$}\!\left/ \!\raisebox{-1ex}{${p}_B(t)$}\right.=\frac{m_A-{m}_A{\left(1-{m}_A-{m}_A\right)}^t}{m_A+{m}_B{\left(1-{m}_A-{m}_A\right)}^t}. $$

(A3)

In the special case of symmetric admixture, m _A  = m _B  = m, the formula is reduced to:

$$ \raisebox{1ex}{${p}_A(t)$}\!\left/ \!\raisebox{-1ex}{${p}_B(t)$}\right.=\frac{1-{\left(1-2\ast m\right)}^t}{1+{\left(1-2\ast m\right)}^t}. $$

(A4)

Dependence of this ratio on generation number and admixture level is demonstrated in Fig. 2.

Fig. 2

Dependence of frequency ratio on population admixture time. The figure shows the expected mutation frequency ratio as a function of admixture time between two populations under different admixture levels ranging from 0.001 to 0.01. The unit of the time scale (X-axis) corresponds to 25 years per generation

The formulae above describe the average dynamics of mutation frequencies in two populations, the ratio deviation caused by genetic drift was not estimated in this simple model.