This document discusses using protein interaction networks as a scaffold to integrate other biological data sources. It describes several high-throughput studies that mapped protein interaction networks in yeast, worm, fruit fly and human. While the protein interaction networks are incomplete and have high false positive rates, topology-based scoring and other filtering methods can be used to identify high-confidence interactions. The document goes on to discuss integrating protein interaction networks with gene expression time series data from yeast to build temporal networks and identify periodically expressed genes and modules involved in cell cycle regulation.