Fig. 1

Overview of the study. a We constructed three whole-genome sequence (WGS) datasets: a general dataset composed of sequences for NA12878 and HG00514; a downsampled NA12878 dataset, and an external dataset with sequences derived from the collected blood samples from three participants. b The reference dataset for the general and downsampled dataset was composed of the structural variation (SV) data for NA128780 and HG00514 available in public datasets. The reference SV data for the external dataset were obtained through paired long-read data-based assembly. c, d For each WGS dataset, we obtained SV results according to the same pipeline and then subjected all SV types’ accuracy, computing resources, and read-depth dependency to fair comparison