As the father of two sons with diagnoses of ASD, I think that it is best to think in terms of an extended spectrum which encompasses 'normal' people. We are all somewhere on this spectrum, but there is a threshold zone around where we cross from "normal" to "diagnosable as ASD". Now, if the distribution of spectral position for males is displaced towards the ASD end of the spectrum relative to females (because of some aspect underlying "normal' maleness, such as testosterone levels) then any genetic variation which pushes someone towards the ASD end of the spectrum by some fixed amount will tend to push more males over the threshold than females.