A new integrative indicator to assess crop genetic diversity? About the publication by Bonneuil et al. (2012), published in Ecological Indicators 23, 280-289

摘要

To study the change of genetic diversity in wheat cultivated varieties over the French territory from the end of the nineteenth century to 2006, Bonneuil et al. (2012) defined and used an indicator to account for the spatial share of the different varieties. However, we found two errors in the implementation of this indicator. The first error is to combine an estimation of weighted genetic diversity among populations with the unweighted Nei coefficient of differentiation among populations (G(ST)). Furthermore, the authors considered, what could be justified, that population varieties cultivated at the end of the nineteenth century and the beginning of the twentieth century, i.e. land races, have now lost their within population diversity due to the process of their maintenance. Then, to retrieve the total genetic diversity present at the period when land races were cultivated, they proposed to add an estimate of the within-variety diversity to the current estimate of the between-variety diversity, which they considered as equal to Nei's parameter (D-ST) of genetic differentiation between populations. This is the second and main error. Indeed, we show that, when each population is reduced to one single line, the expectation of the between-line genetic diversity is not equal to the D-ST at the level of the heterogeneous populations but is near to their total gene diversity. The result of the authors' computation is then a high overestimation of genetic diversity for the period where land races were cultivated. The consequence of the two main errors is that the proposed indicator is not scientifically based and its application leads to erroneous conclusions.