Output:
A list containing six elements: one table–with per-locus counts,
Fisher’s exact test estimates, p-values and true/false columns for
each type of sex-linked loci–and five genlight objects: one with
autosomal loci, and one with each type of sex-linked loci.
Two sets of ‘before’ and ‘after’ plots: one set with female call rate
plotted against male call rate with each data point representing one
locus (one plot before and one plot after removing sex-linked loci
identified by call rate). The other set with proportion of
heterozygous females plotted against proportion of heterozygous males,
and each point representing one locus (one plot before and one after
removing sex-linked loci identified by heterozygosity).
Recommended use: In order to minimize the number of loci
analysed by the function to speed computation time, it is advantageous
to use the filter.sex.linked function after removing of secondary
loci (i.e., those in the same sequenced fragment). This may not be
needed when computation time is not a concern or the number of loci is
small (i.e., ~50,000 SNPs), which will help identify
sex-linked markers in species with short or little-differentiated sex
chromosomes. Additionally, it is strongly recommended to use this
function before other quality filters in order to ensure that (i)
variation in call rate has not been truncated, and (ii) downstream
filtering is done on autosomal loci only. When known-sex individuals are
scarce, we recommend using 15 known-sex individuals of each sex to
identify as many sex-linked loci as possible (even if few), then use
those sex-linked loci to sex all individuals with functioninfer.sex , and then use the new sex assignments to identify the
remaining sex-linked loci with function filter.sex.linked (see
Results section 3).