A long time ago, in the unpublished manuscript of a book on evolutionary biology, I put forward a hypothesis for the origin of blood group antigens that correlates with recent medical findings in relation to the ABO blood group system. My scenario is based upon interactions between invading organisms (whether viral, bacterial, protozoan, or multi-cellular) and their hosts. It is now well known that pathogens and parasites can insert DNA sequences into their hosts' genomes (or otherwise genetically manipulate them) in order to enhance their own survival, and it is becoming more and more obvious that human beings are just as susceptible to these strategies as are other species. I suggest that such genomic reprogramming events could have been the basis of the evolution of blood group antigens, as the latter are disadvantageous to those individuals possessing them: the malaria parasite Plasmodium vivax gains entry to human red blood cells by means of the Duffy antigens, while the presence of all other blood group substances makes it impossible for the host to mount an immune response to the corresponding surface proteins of invading organisms, which the host is forced to treat as ‘self'. Such a scenario gives rise to a completely new perspective, for hitherto it has been assumed that all proteins are of endogenous origin and serve host functions. Furthermore, it has important implications to therapeutic measures, as an alien DNA sequence can be targeted without fear of adverse consequences to the host. In fact, researchers are at present hoping to confer resistance to HIV in some individuals by inactivating the CCR5 gene, which encodes the receptor that HIV uses to enter certain immune cells (1). On the basis of this hypothesis, the Table shows how the existing relationship between the presence of antigens and the respective naturally occurring antibodies in the ABO blood group system can be explained. Significantly, several studies conducted over the last few years correlate with the relative fitnesses suggested in the Table, for they show that, compared with group O individuals, the non-O groups not only have a decreased overall postoperative survival, whatever the type of surgery involved (2), but are also at increased risk of such conditions as coronary heart disease (3). Table Relationship between antigens and antibodies in the ABO blood group system The latter study is of particular relevance to my scenario, as it demonstrates that the increased risk of coronary heart disease for group AB is more than the combined increased risks for groups A and B [ie 23%, 5%, and 11%, respectively] (3). Also relevant is the fact that it is now known that group A is associated with higher levels of low-density lipoprotein (LDL) cholesterol, which can clog up arteries, and group AB is linked to inflammation, which may adversely affect the function of blood vessels, while people with blood type O have higher levels of a compound that has a beneficial effect on blood flow and clotting.