Preimplantation genetic diagnosis (PGD) of single gene disorders relies on PCR-based tests performed on single cells (polar bodies or blastomeres). Despite the use of increasingly robust protocols, allele drop-out (ADO; the failure to amplify one of the two alleles in a heterozygous cell) remains a significant problem for diagnosis using single cell PCR. In extreme cases ADO can affect >40% of amplifications and has already caused several PGD misdiagnoses. We suggest that an improved understanding of the origins of ADO will allow development of more reliable PCR assays. In this study we carefully varied reaction conditions in >3000 single cell amplifications, allowing factors influencing ADO rates to be identified. ADO was found to be affected by amplicon size, amount of DNA degradation, freezing and thawing, the PCR programme, and the number of cells simultaneously amplified. Factors found to have little or no affect on ADO were local DNA sequence, denaturing temperature (94 or 96 degrees C) and cell type. Consideration of the causal factors identified during this study should permit the design of PGD protocols that experience little ADO, thus improving the accuracy of PGD for single gene disorders.