Over the past 20 years, the biotechnology industry has been extraordinarily successful in bringing a wide variety of new products to the market, including recombinant versions of natural proteins such as growth hormone, insulin and the gonadotropins. The availability of the human genome sequence has given us the chance to identify the entire catalogue of human secreted proteins, often called the secretome. One of the challenges of biotechnology research has been to identify the biological activities of these proteins and to identify if any of them could have a therapeutic or pharmacologic use. The paradigm has effectively been reversed, in that it used to be easy to know the biological activity, but difficult to clone, whereas now the contrary is true. Five years on, it is clear that finding new biological activities is a very difficult process. Much of the ground gained in this period has either been through the development of antibodies as therapies or by the use of protein engineering to produce better versions of the proteins that are already being produced.