Dear Editor,
This is a remark on the paper "Unbiased Estimation of Mutation Rates under Fluctuating Final Counts" by Ycart and Veziris, published in your journal in July 2014.
In their paper they made a reanalysis of my study "Drug-Susceptible Mycobacterium tuberculosis Beijing Genotype Does Not Develop Mutation-Conferred Resistance to Rifampin at an Elevated Rate" published in the Journal of Clinical Microbiology in 2003. In my paper I demonstrated, by the use of fluctuation assays, equal capacities of six Beijing and seven non-Beijing strains to mutate to rifampin resistance.

First a brief background to the fluctuation experiment, which is used for mutation rate estimations. Due to the M. tb characteristics as well as the manual lab work, fluctuation assays repeated on a particular M. tb strain will always cause slightly different results between test occasions (internal experiment variation). In our 2003 study, we concluded that the reproducibility of the fluctuation assay, if performed under strictly standardized conditions, seem to be within the power of ten. In other words, the mutation rate for a particular M. tb strain may be 3.5e-8 (i.e. 3.5 mutations per 100 000 000 cell divisions) at test occasion A but 7.1e-8 at test occasion B.
With this in mind I notice two problems in the Ycart et al reanalysis.

Firstly, Ycart and Verziris assert that their re-calculations of the original experimental data show significant differences between the two groups of M. tb genotypes. Noteworthy, their mutation rates still differ less than the power of ten between the two groups of strains but also when compared to the 2003 calculations. It is of significance to remember the before mentioned reproducibility range of the fluctuation experiment and be cautious with definitive conclusions based solely on mathematical calculations.
Secondly, the authors state that they now found a significantly higher mutation rate for the M. tb Beijing strain group (4.37e-8) compared to the non-Beijing group (2.69e-8).
A closer examination of the reanalysis by Ycart et al. shows that the authors mixed up the groups of strains. They say that the strains in the first seven lines in their Table 2 are the Beijing genotype strains. But these are in fact the non-Beijing strains, which is clearly illustrated in the 2003 paper. As a consequence, the calculations by Ycart and Verziris actually lead to an opposite conclusion, i.e. non Beijing strains demonstrate a higher mutation rate. Thereby the reanalyzed data of Ycart et al does not alter the conclusion of the 2003 original study.