probability of two samples sharing a SNP by error

Question

[SemiQuant]

I'm trying to calculate the probability of the same snp occurring in two different samples by chance. I'm pretty lost here, below are the parameters I have, I'm calling SNPs at ≥1% of the reads in targeted, ultradeep sequence data.

gene_length = 500
error_rate = 0.01
read_depth = 100000

[karl.stamm]

What does by chance mean? Do you think errors show up from totally random cosmic rays, or maybe something intrinsic in the instrumentation. If you used the same instrument and reagents and prep method, then systematic errors will occur in many samples 'better than chance'. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3141275/

[SemiQuant]

Thanks. By chance, I just meant that we ignore the systematic error and base the calculation on a random error rate of 0.01 I'll have a read of that reference.

Answer 1

[SemiQuant]

Would this give me the probability of the same error in two samples?

r1 <- error_rate*(read_depth/1000)
(r1*r1)*gene_length
0.05