How many SNPs

Assume \(N_e \approx 10^6\) and \(\mu = 3 \times 10^{-8}\). Then \(\theta=0.12\):

ne=1E6
mu=3E-8
theta=4*ne*mu
theta

[1] 0.12

Watterson’s estimator has us expecting \(S=\theta\sum_{i=1}^{n-1}\frac{1}{i}\) segregating sites. Thus for 10 diploids (20 chromosomes) and 600Mb of sequence*, we expect (in millions of SNPs):

n=20
wfactor=sum(1/1:(n-1))
s=600E6*theta*wfactor
s/1E6

[1] 255.4373

*bp of sequence should be positions where mean quality is > threshold (say 30) and >80% of samples have data.

To try

Filter all sites for quality. Filter all sites for >80% present data.

Filter windows to have >20% data (i.e. 2kb in a 10kb window)