My smidgeon of an opinion:
No. Genes are not created. Mutations happen. Those which bestow reproductive fitness advantage usually increase in frequency. Those which cause non-recessive deleterious conditions are usually purged. Those which cause recessive deleterious conditions usually persist at low frequency. Those which have 'neutral' fitness consequences may be lost or persist, depending on the fickle nature of chance. Modeling how quickly any of that will occur in a real world population is tremendously difficult. But in a nut shell, those are the general patterns.
I believe what you are saying is that populations in Kenya have had to run to catch food, get to the movies, etc for generations. Those that got the most food and started making out in the back row of the theater first, contributed most to the next generation. And from there the cycle repeated itself until people in Kenya were all super fast runners with a soft spot for romantic comedies. Therefore, someone from Kenya would be faster than someone from the U.S. (who, since the dawn of time, have gotten their food from the supermarket) because of generations of 'runner genes'.
However based upon my experiences in most of the world (admittedly not Kenya yet) even the slow, fat, or somewhat ugly people of the world have just as many if not more children, and pass on whatever genetic component there is to being slow, fat, or somewhat ugly. And frankly, most of the world is slow, fat, or somewhat ugly compaired to world class marathoners. There are way more of those genes in Kenya (and everywhere else for that matter) than 'fast runner' genes.
In a nutshell, there is no concentration of athletic specific genes anywhere. Learned skills, sure. As others have said in this tread, sure more Canadians play hockey than Kenyans, but is that due to genetic athletic specialization to hockey? Changes in gene frequency work on much, much slower time lines (99.9999% of the time) than hockey does, should it ever be played professionally again.