drink_phylogenetics_2009010.jpg

Over the holidays Jim Harriman decided to investigate the family tree of alcoholic drinks. To do this, he screen-scraped all the mixed drink recipes he could find online and crunched the recipe data with a program that generates phylogenetic trees, drawing relationships between drink species with matching ingredient genes.

Note that you can make out several different “kingdoms” of drinks after a close look at the tree. I can make out the Gin kingdom, the Orange Juice kingdom, and the Amaretto kingdom, for starters. Then we have the outliers, like a 110 in the Shade, which nobody in his right mind would drink. These are the platypuses and slime molds of the drink world.

I’d love to know how closely this resembles the actual heritage of the recipes in the list. In fact, it would be incredibly cool to do something similar with food recipes. If you processed the ingredient list and preparation details for the world’s apple pies, chicken soups and breads, what cultural information might that hold?

If you want to take a stab at something like this yourself, you can use a free package call PHYLIP to do the computation. It’s the same program used by Jim to create his drink family tree. If you discover anything, make sure to send us a link.

Phylogenetic Tree of Mixed Drinks
PHYLIP – PHYLogeny Inference Package