When we think of the functions of factors that initiate tumour development, we typically think of classical oncogenes and tumour suppressors that regulate cellular activities such as proliferation and death. But gene-expression analysis and human genome studies have recently shown us that cancer is also associated with gene products and cell activities that we hadn't considered before.

Two articles in this issue indicate that it might be time to increase the scope of what types of proteins and cell activities we consider to be tumorigenic. On page 179, Davide Ruggero and Pier Paolo Pandolfi summarize a number of proteins that directly affect ribosome biogenesis and protein translation. Increased protein synthesis might not simply be a downstream effect of a cell proliferating out of control, but a contributing factor to transformation.

Charis Eng and colleagues explain on page 193 how mutations in mitochondrial enzymes that regulate the Krebs tricarboxylic-acid cycle can lead to a number of inherited neoplasia syndromes. In addition to the ability of mitochondrial membranes to control the release of pro-apoptotic proteins, their role in regulating energy production can also affect tumorigenesis. Even the classical definition of what was considered to be a tumour suppressor has been modified, as these proteins can be subdivided into 'gatekeepers' and 'caretakers'. On page 169, Ian Hickson explains how defects in caretakers can perturb genomic stability and thereby facilitate tumour development, even though they do not directly induce transformation.

Studies such as these remind us that we will have to round up more than just the usual suspects to learn how tumours form, grow and spread — it is time to look for new oncogenes and tumour suppressors in places that we haven't looked before.