Although drug resistance is commonly used as an indicator of gene transfer in various cellular contexts, the assessment of drug resistance is often imprecise and over-estimated. To measure accurately transduction efficiencies of the retroviral-mediated transfer of genes encoding the neomycine phosphotransferase (Neo(r)) and porcine major histocompatibility (MHC) class II in pig bone marrow cells (BMC), the fraction of targeted progenitors was evaluated by both colony-forming unit granulocytes/macrophages assays (G418r CFU-GM) and by PCR analysis of the transgenes (Tg). Transduced and untransduced BMC were selected at different concentrations of G418 and revealed high individual variability of drug sensitivity. Comparison of the results obtained by estimating the CFU frequency and the PCR assays on drug-resistant colonies demonstrated a marked overestimation of BM transduction rates when determined by G418 resistance alone, because only approximately one-third of individual colonies were positive for both the Neo(r) and the class II Tg. Because this discrepancy is likely to affect the overall assessment of transduction rates using drug resistance markers, our data attest for the need of a combination of molecular assays to determine transduction efficiencies accurately.