USC Algorithm Helps Researchers Cut NGS Costs by Predicting Needed Sequencing Depth