By contrast, this new enzyme, developed by Burckhard Seelig's lab at UM's College of Biological Sciences, was developed in the same way enzymes evolve in nature. A large quantity of candidate proteins were placed together in culture and screened with every successive generation for their ability to perform a desired function (in this case, joining two pieces of RNA together). Unlike rational enzyme design, this approach isn't limited by what the researchers know about enzyme structure. All the researchers really need to know is what they want from the enzyme. Evolution finds the best way to get there.