NCI-funded investigators performed a comprehensive study of 45 human large gene and protein networks, or “interactomes,” thereby providing a reliable guide to help researchers like you select the most suitable network for your specific needs. The evaluation pipeline, a practical tool derived from this research, is available to help you assess and integrate the best interactomes for your work, enabling decision-making in your research.

Navigating the extensive landscape of interactomes for cancer research can be challenging. But with the rapid advancement of genomic and proteomic technologies, numerous interactomes are now available, each with its own strengths and limitations. The researchers found that larger composite networks—such as HumanNet, STRING, and FunCoup—are particularly effective at identifying disease genes, including those associated with cancer. In contrast, smaller networks—like DIP, Reactome, and SIGNOR—perform better in predicting molecular interactions.

Dr. Trey Ideker, corresponding author, stated, “This research provides the most expansive and up-to-date reference for network selection. Unlike previous studies that benchmarked only a few human interactomes, this research expands the scope by examining a wide range of networks. Our approach assesses current interactomes and offers a framework for evaluating emerging and updated interaction networks in the future.”

NCI’s Informatics Technology for Cancer Research Program supported this work.