We found that the mean coverage and read coverage distributions for the two sets of SNPs is virtually identical.
The upper panel: whole-genome sequencing set; the lower panel: whole-exome sequencing set; lfet: the box plot depicts the distribution of mean coverage; right: the box plot depicts the distribution of fraction of bases covered by at least 1 reads, 4 reads, 10 reads and 20 reads across the sequencing samples.
At 10× mean coverage and 0.5% error, the distribution of correctly formed loci matched closely that of the true distribution, differing only by having somewhat fewer loci (a 14% reduction) than actually exist.
We examined the distribution of mean coverage of the captured reads mapping to the 500 targeted gene pairs: more than 400 gene pairs (407 and 404 for Maxxa and TX2094, respectively) are represented by a mean read depth of more than 20X, which was our goal for identifying homeologous copies and alleles in allopolyploid cottons.
The interquartile ranges (IQR) of the control bulk and high bulk of green seedling in response to salt are 5792 and 7105, respectively, indicating a highly nonuniform distribution compared to IQR = 119 for a Poisson distribution when the mean coverage is 7800×.
The central peak of the distribution indicated a mean coverage of 41-43.
In theory, this assumption entails an equality between the mean and the variance of the coverage distribution.
Similarly, IQR of low and high bulk of seed size are 4321 and 4318, compared with IQR = 108 for a Poisson distribution when the mean coverage is 6440×.
Using the eight SSC samples, we investigated the coverage behaviors of the WGS and WES data by the following: distribution of the read depth, mean coverage, coverage fraction at X reads, coefficient of variation (C v ) (See Methods).
The negative binomial distribution is described by the mean coverage (μ cov ) and a size parameter (α cov ) reflecting the overdispersion.
