Researchers at Scripps Translational Science Institute and the University of California, San Diego, found that creating two libraries with different insert sizes was the best way to detect structural variation, while the amount of sequencing needed to detect any given transcript follows a distribution curve.