The aim of this study was to screen for key biomarkers of osteosarcoma (OS) by tracking altered modules. Protein-protein interaction (PPI) networks of OS and normal groups were constructed and re-weighted using the Pearson correlation coefficient (PCC), respectively. The condition-specific modules were explored from OS and normal PPI networks using a clique-merging algorithm. Altered modules were identified by a maximum weight bipartite-matching method. The important biological pathways in OS were identified by a pathway-enrichment analysis using genes from disrupted modules.