The quality of the genotyping data at a specific marker was controlled using the "drop locus" command.
We manually set 2 thresholds on the cluster mean (represented as gray lines) at 1.5 (dropping loci with low count cluster greater than 1.5) and 2 (dropping loci with high count cluster less than 2).
b: same as a for G libraries, we dropped loci 8, 9, 12, 22 (standard deviation greater than 0.5).
The reasons for dropping loci were either lower separation between read count clusters or read alignment to multiple locations in the genome.
As a result, we dropped locus 53 (low read count cluster with mean greater than 1.5) and loci 39, 62, 63, 80 and 99 (high read count cluster with mean less than 2).
We manually set thresholds (gray lines) at 3 (mean) and 0.5 (standard deviation) and we dropped locus 5 (low read count cluster had mean greater than 3), and loci 10 and 15 (standard deviation greater than 0.5).
Larger cutoff values tend to remove the background noise of the test statistics by dropping more loci from a window.
However, within that site, occupancy was highest close to the telomere end and progressively dropped at loci towards the centromere (Figure 6C).
We manually set thresholds (represented as gray lines) at 2 (mean) and 0.6 (standard deviation), which dropped out loci 13, 17, 19, 39, 62, 80, 82, 93, 96 and 99 (high read count cluster had mean less than 2) and locus 53 (standard deviation greater than 0.6).
We dropped this locus from the dataset, recalculated all estimates of genetic diversity and Ne and compared values between original and truncated datasets.
