Abstract—Several structural variation identification approaches have been developed with various considerations and merits. As a scientist, it is important to choose the suitable tool. Unfortunately, there is a lack of gold standards to benchmark these approaches. A Structural Variation Benchmarking and Evaluation Tool (SV-BET) is proposed, which composed of three main components: benchmark creator, mapper, and evaluator. We use the proposed tool to evaluate the performances of seven available approaches: Pindel, Prism, Delly, Softsearch, Softsv, Socrates, and Manta. These tools are based on split-read-based approaches for detecting structural variations. SVBET is tested using the Escherichia coli K12 genome. Results show the sensitivity and positive-predictive value of detected structural variations and breakpoints for each approach. SV-BET can be used to evaluate the performance of other SV identification algorithms.