TD-TSPS: A hybrid strategy method for TD detection based on two-step progressive segmentation

Tandem duplication (TD) represents a crucial type of structural variations within the human genome. When the sequencing depth is low, TD signal at each single nucleotide position becomes more indistinct. So the detection of TDs under low coverage remains a challenging task. This paper proposes a method called TD-TSPS (Two-Step Progressive Segmentation for TD detection) on whole genome sequencing data. A two-step progressive segmentation strategy is employed to divide the genome into continuous and similar bins. Additionally, it integrates split read and paired-end mapping strategies to refine TD regions. Performance tests on simulated and real datasets show that TD-TSPS achieves a good F1-score. Therefore, it can be used as an effective tool for TDs detection.