SUMMARY.Currently, five platelet alloantigen (alloAg) systems have been established (HPA‐1, ‐2, ‐3, ‐4, ‐5). Three of these are expressed on the glycoprotein (GP) IIb‐IIIa complex, HPA‐1, HPA‐3 and HPA‐4, inherited in an autosomal codominant mode. Recent investigations of the molecular basis of these platelet alloantigen systems have shown that only one nucleic acid base substitution in the genes encoding for GP IIb and GP IIIa is responsible for the polymorphism. This substitution is reflected in a difference in restriction enzyme recognition allowing platelet alloantigen typing by restriction fragment length polymorphism (RFLP) analysis of DNA amplified by the polymerase chain reaction (PCR). To validate the PCR technology for platelet typing, we have compared PCR‐RFLP with monoclonal‐antibody‐specific immobilization of platelet antigens (MAIPA). For this purpose, we have studied different Glanzmann thrombasthenic families and particularly heterozygous individuals, who are not lacking GP IIb‐IIIa, as a model to detect the occurence of discrepancies between these two technologies. In two families, we have found differences between molecular biology and serological methods with the lack of expression of one antigen on the platelet membrane surface. In the first family, the abnormality is related to the HPA‐1 alloantigen system with three informative members; in the second, the HPA‐3 alloantigen system is concerned with two informative members. Considering these results, there may not always be a perfect correlation between molecular biology and serological methods, as an unknown molecular defect could interfere with the PCR results and lead to false platelet typing. In cases of suspected post‐transfusion purpura (PTP) or fetuses at risk for neonatal alloimmune thrombocytopenia (NAITP), great care must be taken in the interpretation of the results if on