The big question is: did the present-day high allele frequency in Europeans happen because of natural selection or because of admixture with a population that was already lactase persistent?
For example, the lactase persistence allele occurs at a non-trivial frequency in present-day inhabitants of the Americas, whereas it was zero there a few thousand years ago, with the culprit being post-1492 European colonization. The frequency change in the Americas didn't happen because of natural selection, but because a new population (Europeans) moved in.
If admixture with a lactase persistent population L is at play, then the question remains how L became lactase persistent in the first place. However, this transforms the problem from (a) seeking something in the European cultural or natural environment acting as an agent of selection, into (b) seeking something in the cultural/natural environment of population L. I don't know what L might be, but seeking a population with lots of cows is a good place to start...
Mol Biol Evol (2014)
Direct estimates of natural selection in Iberia indicate calcium absorption was not the only driver of lactase persistence in Europe
Oddný Ósk Sverrisdóttir et al.
Lactase persistence (LP) is a genetically determined trait whereby the enzyme lactase is expressed throughout adult life. Lactase is necessary for the digestion of lactose – the main carbohydrate in milk – and its production is down-regulated after the weaning period in most humans and all other mammals studied. Several sources of evidence indicate that LP has evolved independently, in different parts of the world over the last 10,000 years, and has been subject to strong natural selection in dairying populations. In Europeans LP is strongly associated with, and probably caused by, a single C to T mutation 13,910bp upstream of the lactase (LCT) gene (-13,910*T). Despite a considerable body of research, the reasons why LP should provide such a strong selective advantage remains poorly understood. In this study we examine one of the most widely cited hypotheses for selection on LP – that fresh milk consumption supplements the poor vitamin D and calcium status of northern Europe's early farmers (the calcium assimilation hypothesis). We do this by testing for natural selection on -13,910*T using ancient DNA data from the skeletal remains of eight late Neolithic Iberian individuals, whom we would not expect to have poor vitamin D and calcium status because of relatively high incident UVB-light levels. None of the 8 samples successfully typed in the study had the derived T-allele. In addition, we reanalyse published data from French Neolithic remains to both test for population continuity and further examine the evolution of LP in the region. Using simulations that accommodate genetic drift, natural selection, uncertainty in calibrated radiocarbon dates, and sampling error, we find that natural selection is still required to explain the observed increase in allele frequency. We conclude that the calcium assimilation hypothesis is insufficient to explain the spread of lactase persistence in Europe.