A fundamental goal of systems biology is to define a biological system precisely, such that it becomes possible to predict the outcome of perturbing that system. Yesterday (Jan. 22) a team of researchers from German drug discovery firm Cellzome and the European Molecular Biology Laboratory reported in Nature a significant step toward the creation of such models, at least in budding yeast.

Giulio Superti-Furga and Robert Russell, and colleagues systematically tagged 6,466 Saccharomyces cerevisiae ORFs at the C-terminus with a "tandem-affinity-purification"(TAP) tag. Of these, they were able to grow 3,206 proteins and purify 1,993 of them. They analyzed these proteins using mass spectrometry, identifying some 491 protein complexes, of which 257 are novel.

This study fleshes out the budding yeast protein interaction map the team first produced almost exactly four years ago -- especially membrane-associated proteins. The authors used a new protocol to improve their yield of "this notoriously more difficult and challenging class of protein"; of 628 tagged membrane proteins, the team successfully purified 54% (340) compared to 62% for soluble proteins. (In contrast, the team's 2002 study purified just 40 membrane-associated proteins.)

But this article goes much further than creating an expanded "parts list" for the organism. The authors subdivide their complexes into cores and modules, with cores being, well, core and relatively constant, and modules being small multiprotein complexes that can modify a core's behavior. "Modularity might very well represent a general attribute of living matter," the authors write, "with de novo invention being rare and reuse the norm."

A small number of cores could thus facilitate a large number of functions by virtue of swapping accessory modules, the authors note. If true, that knowledge could aid work of systems biologists like those at the Molecular Sciences Institute, whose Alpha Project is an attempt to model the yeast mating response in silico, by helping them develop more accurate simulations.