Authors:	Bork P
CellNetworks People:	Bork Peer
Journal:	Cell 144, March 18, 2011

The amount of data being produced to study biological systems at various scales, from molecules to ecosystems, is growing exponentially, and handling this data from local production to its storage in publicly accessible and integratable depositories poses technicalchallenges that some areas of biology are already confronting. But the conceptual challenges ahead may be even more daunting. Amajor one is quantifying
the impact of context, that is, experimental constraints and environmental factors that influence results. Internal and environmental properties together characterize biological systems, exemplified by human diseases, which are affected by complex genetic and environmental components, the latter being barely understood and still frequently neglected in current studies. Another challenge is how much we can abstract from observations derived from cultivated cell lines, given the absence of a native tissue context? Many current technologies impose noise onto real biological signals (for instance, studying cell populations rather than individual cells is frequently unavoidable), and given the complexity of biological systems in terms of their many interacting elements
and confounding variables, how are we to estimate which aspects of a finding remain valid in other settings? Thus, there is an urgent need for generalized formal descriptions
of the state and the environmental context of biological systems (metadata), which would not only improve the reproducibility and comparability of observations, but would also enable strategies for quantifying the impact of environmental conditions. Such efforts will help to minimize
data overinterpretation (as can easily occurwith indirect correlations) and reduce the accumulation of misleading results.