A population genetic analysis of male homosexuality

As a companion to the previous post, I pass along this article by Ciani et al. arguing that only a two-locus genetic model for male homosexuality with at least one locus on the X chromosome, in which gene expression is sexually antagonistic (increasing female fitness but decreasing male fitness), accounts for all the known empirical data. That data is interesting (as described in this account in Slate):

It starts with four curious patterns. First, male homosexuality occurs at a low but stable frequency in a wide range of societies. Second, the female relatives of gay men produce children at a higher rate than other women do. Third, among these female relatives, those related to the gay man's mother produce children at a higher rate than do those related to his father. Fourth, among the man's male relatives, homosexuality is more common in those related to his mother than in those related to his father.