"These epi-marks protect fathers and mothers from excess or underexposure to testosterone — when they carry over to opposite-sex offspring, it can cause the masculinization of females or the feminization of males," Rice says, which can lead to a child becoming gay. Rice notes that these markers are "highly variable" and that only strong epi-marks will result in a homosexual offspring.
I'm not buying this masculinization of females/feminizing of males stuff. I know far too many people of all genders and all sexual orientations who don't fit the stereotypes.
Not saying that there isn't a heritable component to sexual preference and gender identity, just saying that the way this is reported sounds way too schematic to be believable.
Also, the reporter, if not the researchers, thoroughly confuse gender identity (male/female) with sexual preference (choice of gender in a sexual partner). The two are not necessarily paired together. Furthermore, both male/female gender identity and gay/straight preferences are false dichotomies: both are continuums. Both identity and preference can be quite fluid.