Amplify the gene in question and use a quantitative realtime PCR assay and look for differences in high resolution melt curve vs control.
Or, drink 3 glasses of milk, wait a couple months on lactose free diet, and drink some of the lactose-free stuff. Then analyse your feelings? That is the old school way, the gut feeling. That and tallying those in your lineage who have it, and arriving at a conclusion that is, if not empirical, at least probable in a Bayesian sense.
I would recommend the 23andme sequencing, I got mine done. As long as you are comfortable with the reported error rate- which could be diminished with repeat sequencing and some sort of alignment algorithm. Otherwise you are talking a lot of work and access to expensive hardware.