They actually did something kinda clever. They were able to calculate/model a false positive rate by looking at the homozygosity of the X chromosome. Since the

They actually did something kinda clever. They were able to calculate/model a false positive rate by looking at the homozygosity of the X chromosome. Since the samples were male, there should only be one X, so any heterozygosity would have to be errors. Not saying this is an absolute correct way of going about this (no one really knows), but it shows they did at least thought about this issue.