BACKGROUND: Cleft lip with or without cleft palate (CL(P)) and isolated cleft palate (CP) are separate clinical entities and for both polygenic multifactorial aetiology has been proposed. Parents of children with orofacial clefting have been shown to have distinctive differences in their facial shape when compared to matched controls. OBJECTIVE: To test the hypothesis that genetic and morphometric factors predispose to orofacial clefting and that these markers differ for CL(P) and CP. Methods-Polymorphisms at the transforming growth factor alpha (TGFalpha) locus in 83 parents of children with nonsyndromic orofacial clefts were analysed, and their craniofacial morphology was assessed using lateral cephalometry. RESULTS: Parents of children with CL(P) and CP showed an increased frequency of the TGFalpha/TaqI C2 allele (RR=4.10, p=0.009) relative to the comparison group. Also the TGFalpha/BamHI A1 allele was more prevalent in the CP parents. MULTIVARIATE STATISTICAL ANALYSIS: Using stepwise logistic regression analysis the TGFalpha/TaqI C2 polymorphism provides the best model for liability to orofacial clefting. To determine the type of clefting a model involving interaction between the parental TGFalpha/BamHI and TGFalpha/RsaI genotypes showed the best fit. Using genotype only to predict the clefting defect in the children according to parental genotype, 68.3% could be correctly classified. By adding information on craniofacial measurements in the parents, 76% of CP and 94% of CL(P) parents could be correctly classified. CONCLUSIONS: This study provides a model for prediction of liability to orofacial clefting. These findings suggest that different molecular aberrations at the TGFalpha locus may modify the risk for CP and CL(P).