If you are investigating low-frequency and rare variants for a complex trait using exome sequencing, why would one consider using different populations (African ancestry and European ancestry) especially considering that the populations have different allele frequencies. Is the main object to see if they're could possibly be novel rare variants in each population at different allele frequencies?
Too long for a comment, but sort of:
Mainly you do this because these populations have segregated long time ago and developed differently, so it is more likely to identify these polymorphisms. It is also helpful to use populations for this which mixed not too much with other populations (thats why north-americans are usually not used here, as America was kind of a melting pot).
You will find rare SNPs for pigmentation in europeans that are not present in the more ancient african population (and must therefore be occured after the populations divided), examples are the so called "golden mutation" in SLC24A5 (rs1426654) or a polymorphism in IRF4 (rs12203592).