Mouse models with targeted disruptions of coagulation factor genes are used to study disorders of haemostasis such as haemophilia. Standard protocols to obtain biopsies for genotyping in breeding programmes are based on invasive sampling methods such as tail clipping. These procedures imply a high risk of fatal bleeding, especially in haemophilic mouse models. Here we used a non-invasive sampling method obtaining stool samples for DNA isolation in a breeding programme, aiming to introduce targeted disruptions of Fc receptor genes (Fc gamma receptor IIB and III) into the haemophilia A mouse model (factor VIII deficiency). Faecal pellets were reliably obtained from individual mice and high-quality DNA was extracted with a mean yield of 7.1 microg/pellet. Polymerase chain reaction amplification of wild-type and knockout alleles for Fc receptor and factor VIII genes was similar, comparing stool and peripheral blood as the source of genomic DNA. Definite genotype identification was achieved in a first attempt in 336 of 352 analyses (95%). Repeated analysis of homozygous knockout animals confirmed the first result in all cases. No animal was lost due to bleeding from the procedure. In conclusion, DNA isolation from stool is a preferable method for genotyping in laboratory animals, especially in models of bleeding disorders. Avoiding loss of animals due to bleeding implies a substantial improvement in animal welfare by reducing the number of animals used and may also advance the effectiveness of breeding programmes in these disease models.