Coincident with the non-invasive collection of samples from free-ranging, unmarked animals come issues of sample anonymity. Weddell seals are an ecologically important Antarctic species, but their distribution overlaps humans and their life-history behaviours limit direct observation. Faecal deposits on sea ice provide a valuable, non-invasive window into their physiology and health, but meaningful interpretation hinges on determining sample sex. Historically, sex-typing of faecal samples has relied on polymerase chain reaction (PCR) to detect sex-specific genetic markers. An alternative approach involves quantifying faecal steroid metabolites (FSM), which may offer a more cost-effective method while also yielding broader physiological insights. In this study, we used the standard PCR approach to establish sample sex and then evaluated the ability of FSMs to discriminate between sexes. PCR of the SRY and ZFX/ZFY loci successfully sex-typed 96% of the samples collected. Of the range of predictive models tested, the generalised linear model (glm) with the lowest AIC value, that included the main effects of untransformed steroid metabolite levels, predicted sample sex from EIA of FSM with sex established by PCR in 81 of 89 samples tested (91% accuracy). Log transformations of steroid levels and increased model complexity only marginally improved prediction accuracy to 93%. In conjunction with understanding the broader endocrinology of a species, the quantification of steroid metabolites by EIA serves as an alternative means of accurately sex-typing anonymous faecal samples across an increasing range of vertebrate taxa.
HOFF et al. (Sat,) studied this question.