A recent edition of radio programme KERA Think has a fantastic discussion on development and the neuroscience of sexual attraction in its many forms.
The programme is a discussion with Simon LeVay, a neuroscientist who raised a lot of eyebrows by finding differences in the brain structure of gay and straight men in a 1991 study.
The science has massively advanced since then and LeVay gives a fascinating and lucid account of what know about the biology of the rainbow of sexual attraction – and where the mysteries still lie.
Mind Hacks 
On page 210 of his book, LeVay discusses my model for the development of heterosexual and homosexual preferences: “This model is attractive in that it solves the “binding problem” of sexual attraction. By that I mean the problem of why all the different features of men or women (visual appearance and feel of face, body, and genitals; voice quality, smell; personality and behavior, etc.) attract people as a more or less coherent package representing one sex, rather than as an arbitrary collage of male and female characteristics. If all these characteristics come to be attractive because they were experienced in association with a male- or female-specific pheromone, then they will naturally go together even in the absence of complex genetically coded instructions.”
—
We now know considerable more about the complexity of genetically coded instructions — enough to say that what is known is insufficient in attempts to explain anything about the adaptive evolution of sexual preferences.
LeVay adds that “Still, even in fruit flies, other sensory input besides pheromones — acoustic, tactile, and visual stimuli — play a role in sexual attraction, and sex specific responses to these stimuli appear to be innate rather than learned by association [36.]. We simply don’t know where the boundary between prespecified attraction and learned association lie in our own species, nor do we have compelling evidence for the primacy of one sense over another.”
—
Neuroscientists have known for many years precisely where the boundary lies between genetically predisposed attraction and learned associations in species from microbes to man. The most compelling evidence available shows that the epigenetic effects of olfactory/pheromonal input clearly establish olfaction as the primary sense and that it is responsible for adaptive evolution via ecological, social, neurogenic, and socio-cognitive niche construction across the evolutionary continuum that includes differences in sexual orientation.
See for example: Kohl, J.V. (2012) Human pheromones and food odors: epigenetic influences on the socioaffective nature of evolved behaviors. Socioaffective Neuroscience & Psychology, 2: 17338. http://dx.doi.org/10.3402/snp.v2i0.17338 (10 pages)
As an alternative read the 57-page journal article concurrently published as a book chapter (that LeVay discussed in the context of my model): The Mind’s Eyes: Human pheromones, neuroscience, and male sexual preferences http://www2.hu-berlin.de/sexology/BIB/kohl.htm