Fig. 2From: InvBFM: finding genomic inversions from high-throughput sequence data based on feature miningWorkflow of InvBFM. It includes two major parts: (i) Training model. Bench-mark file is used to locate true inversion regions and non-SV regions, and then InvBFM extracts features from sequence reads around each label region to train a classification model by SVM. (ii) Calling inversion. Results of several tools are integrated as candidate inversion sites, then InvBFM extracts the same set of sequence features from sequence reads and calls inversion using the trained classification modelBack to article page