Identification of sex reversal
Sex linkage of the 3 markers was tested on 125 laboratory-raised froglets. Individual molecular sexing was performed independently with each marker, and we subsequently checked if the identified genotype matched with the phenotypic sex. For each marker, we calculated the rates of female-to-male and male-to-female sex reversal, i.e. the proportion of phenotypic males among genetic females (XX) and the proportion of phenotypic females among genetic males (XY), respectively, as well as the proportion of sex-reversed individuals within each phenotypic sex. If a marker indicated sex reversal for a laboratory-raised individual, we accepted the result only if a second DNA sample extracted from the other stored foot of that individual gave the same result as the first one (i.e. to avoid false identifications of sex reversals due to human error during the molecular laboratory work). Because the second DNA sample always confirmed the assumptions from the first one, we found these genotypes to be unambiguous. Note that the 125 froglets came from an experiment in which their siblings were exposed to various treatments (Bókony et al., 2020); here we used some of the genetic sex data of those treated siblings to evaluate whether our findings of sex-reversed froglets may have been due to null alleles, sex-chromosome recombination or mutation, or being sired by a sex-reversed parent.
Because the above analyses showed that one of our three markers (Rds2) would not be suitable for sexing in our populations (see Results), we used the other two markers to evaluate two sexing methods. In Method 1, we screened all laboratory-raised individuals for the marker with the highest sex linkage (Rds3) and we accepted an individual to be normal male or female if its Rds3 genotype was in accordance with its phenotypic sex. Those individuals that seemed to be sex-reversed by Rds3 were screened for the marker with the second highest sex linkage (Rds1) as well and were accepted to be sex-reversed only if both markers confirmed sex reversal. In Method 2, all laboratory-raised individuals were screened for both Rds1 and Rds3 and genotyping was considered to be successful only if both markers gave the same result. In both methods, individuals with discrepant genotyping results were considered to be of unknown genetic sex.
We estimated sex-reversal frequency in the wild-captured adults using Method 1, because this sexing method performed best in the laboratory-raised individuals (see Results). As females are more difficult to find and capture than males, the majority of the investigated adults were males, so we had too few females to provide a reliable estimate of female-to-male sex-reversal rate in adults. Therefore, we report the proportion of sex-reversed individuals (XX males) among the phenotypic males (hereafter referred to as XX/male ratio) as a measure of female-to-male sex-reversal frequency.