Hadn't heard that before - could it not be someone else that cut themselves and wiped it on the sacred shroud whilst no one was looking?
But does O really originate in Africa? Or for that matter, is there really anywhere that can be ascribed to be the origin of O (not AB, A or B) or A or B for that matter? Or that A and B are later developments? It seems to me that disease that selectively attacks people with certain blood groups could account for the often rather arbitrary distribution of blood type. It could be the case that all populations had every type but certain diseases wiped out or severely depleted those people with certain blood types, A or B for example. Besides the ABO system is rather over-simplified when it comes to an actual analysis of blood. It serves for medicine but not necessarily as a way to understand evolution, or origins.
Just to turn things around a bit try O as less opportunistic for diseases than either A and/or B (meaning in this case AB/A/B). Diseases learn to attack A and/or B and selectively wipe out people with A and/or B. Leaving people without A and/or B, that is O, as dominant or increased relatively in the population. So, where there has been a lot of disease one might expect O to be dominant, but where disease has been less one might expect A and/or B to still exist. One could perhaps correlate this with zones of the planet - areas where diseases can thrive would be more likely to be O dominant. But, of course, diseases can be somewhat arbitrary and conditions for their success can vary not only in the long term but in the short term. Epidemics spread and then burn themselves out due to geographical or man-made barriers. Weather patterns and temperatures of areas change. Opportunities rise and fall. So rather than explaining migration, the ABO system might explain the demographic of some diseases, or, indeed, of variations in climate.
But, for me, the really interesting part of this theory is what it means when compared with climates, time and civilization. For, if areas or zones of relatively cooler climates have a high prevalence of O, then, given that the climates have been relatively stable for many millennia, these populations may well be the oldest of all. Of course, this sort of thing depends on a great many factors, as I am sure you will realise. But you ought to admit, the notion is indeed rather provocative when one looks at the demographic of blood types world wide.