Hybrid classification: Insights from genetic map-based studies of experimental hybrids

Correct interpretations of hybridization phenomena rely on accurate methods of detecting hybrids and estimating genotypic classes. Multilocus molecula r data sets have proved to be particularly useful for hybrid identification and genealogical reconstruction. One widely recognized limitation to this approach is that hybrid and parental genotypic classes often differ minimal ly in terms of expected marker numbers, and an extremely large number of mo lecular markers are required to distinguish between them. A more fundamenta l difficulty, and the focus of this study, concerns the potential for selec tion to bias marker numbers in hybrids, thus leading to faulty genealogical assignments. In this paper, we compare observed and expected marker number s in 170 hybrids of known pedigree between two wild sunflower species, Heli anthus annuus and H. petiolaris. All 170 hybrid plants had fewer H. petiola ris markers than the expected means. In fact, with the exception of two ind ividuals, marker numbers were more consistent with pedigrees involving thre e to seven backcrossed generations, rather than the two generations of back crossing actually employed. Although the implications of these data for gen ealogical reconstruction are discouraging, it is the actual genetic constit ution of hybrids, not the pedigree, that is most predictive of their charac teristics or behavior. Thus, we suggest that hybrids be classified either i n terms of genetic relatedness or in terms of the admixture of parental-spe cies genes they carry. Both represent simple approaches for generating biol ogically cohesive categories for ecological or comparative studies.