We have become used to considering more than one gene at a time, so why not consider more than one trait at a time? And why not add in some environmental variation too? Yes, it's complicated, but two Reviews in this month's issue discuss why this should be the future for both plant and animal research.

In plants, linking traits to genes has been given a recent boost by the development of tools for genome-wide association studies (discussed in a Review by Bergelson and Roux on p867 for Arabidopsis thaliana and in a Research Highlight on p816 for rice). In an agricultural context, identifying genes that underlie phenotypic variation can have enormous benefits for improving crop quality and yield. But the ecological conditions that plants are exposed to around the world vary tremendously. Bergelson and Roux argue that greater benefit will be derived if ecology is added to the mix, with plants being phenotyped under different natural conditions.

However, measuring one trait in several conditions only takes us part way to understanding adaptation, as there can be trade-offs and interactions among traits and genes can have pleiotropic effects. Thus, argue David Houle and colleagues (Review, p855), we need a serious drive to develop phenomics. Phenomics gathers high-dimensional phenotype data on an organism-wide scale and, unsurprisingly given the scale of this challenge, lags far behind genomics. But the idea is gaining momentum and several collaborative projects in species from plants to humans are underway.

The authors of these articles make no pretence that scaling genotype–phenotype analyses to multiple traits and environments will be simple. But they argue that this is where we should channel our efforts.