Unbiased population heterozygosity estimates from genome?wide sequence data

Heterozygosity is a metric of genetic variability frequently used to inform the management threatened taxa. Estimating observed and expected heterozygosities from genome-wide sequence data has become increasingly common, these estimates are often derived directly genotypes at single nucleotide polymorphism (SNP) markers. While many SNP markers can provide precise processes, results ‘downstream’ analysis with may depend heavily on ‘upstream’ filtering decisions. Here we explore downstream consequences sample size, rare allele filtering, missing thresholds known population structure heterozygosity using two reduced-representation sequencing datasets, one mosquito Aedes aegypti (ddRADseq) other grasshopper, Keyacris scurra (DArTseq). We show that based polymorphic only (i.e. heterozygosity) always biased by global size (N), smaller N producing larger estimates. By contrast, unbiased when calculations consider monomorphic as well information or autosomal heterozygosity). also differentiated populations analysed together while remains unbiased. sites included, diverge in proportion amount permitted each site. make three recommendations for estimating heterozygosity: (a) should be reported instead (or addition to) heterozygosity; (b) any omitted (c) independent runs. This facilitate comparisons within across studies between measures heterozygosity.