If you want, try doing a 'bychr' Weac run to view individual chromosomes. You may see something interesting with the fluctuations of admixture percentages.
The Moorjani et al. 2.4% is a mean with a standard deviation of 0.3%, additionally I would think that the Sub-Saharan they refer to includes all African snps, irrespective of whether they may be found elsewhere in high frequencies. They mislead the reader somewhat but the point is still the same, they simply overemphasized the African element a bit. I'm not concerned about this.