The next large-scale human genome project after HapMap should catalogue inherited variation in the general population that directly affects gene function, argues Richard Gibbs.