Fig. 1

Differences in skin microbial composition (beta diversity) between the poorly- and highly-attractive groups. Each participant’s attractiveness to mosquitoes was measured using a bioassay, and differences in microbial composition between participants calculated using centralised log ratio transformed, DEICODE beta diversities from 16S rRNA data (see Methods). The sparse partial least squares discriminant analysis (sPLS-DA) sample plot shown was used to compare differences in microbial composition between the poorly-attractive group (orange) and highly-attractive group (blue). Individuals are presented by small triangles (poorly-attractive) or small circles (highly-attractive). Data were scaled (centred and standardised). The centroids (stars) represent the mean microbial composition on the first and second components for each group. The ellipse plots (large circles) represent 95% confidence intervals for the relative attractiveness groups