However, we should note that, the reference variant data from the 1000 Genomes Project and the Exome Sequencing Project were produced by different platforms, most of which were next generation sequencing platforms. The sequencing depth, coverage and data analysis pipelines might affect the variant detection rate. It is the consideration that the variant number from different platforms might not be compared directly. So we focused on the locations of the rare variants on the protein, and the analysis strategy is feasible in our study. More importantly, in our in vitro assays, three private variants were shown to alter the ability of DLC1 to inhibit cell migration or the subcellular localization of the protein, which supported the notion that private variants might also play major roles in the pathological process of complex diseases. In addition, the extended N-terminal region of DLC1 isoform 1 harbors 83% of the private variants identified in the CHD cohort in a non-random manner. The relatively high transcriptional level of DLC1 isoform 1 in human heart tissues implies that the unique N-terminal region may possess a tissue-specific function in the cardiovascular system. However, future studies are necessary to elucidate the details. Cell migration is an evolutionarily conserved mechanism that includes four steps: polarization, protrusion, adhesion and retraction. Actin is primarily involved in the last three steps. Studies have confirmed that DLC1 can function in the regulation of actin cytoskeletal organization and cell migration, suggesting that DLC1 acts as an important regulator of migration. It is essential for endothelial cells in the outflow tract and atrioventricular regions to migrate into the cardiac jelly during embryonic heart development. Similarly, the migration of cardiac neural crest cells is also a crucial event during heart development, and the inappropriate timing or path of cardiac neural crest cell migration will cause cardiac congenital Torin 1 anomalies. Thus, if the migration regulatory ability of DLC1 is impaired in the early stage of fetal cardiac development, it is reasonable to speculate that inaccurate developmental consequences, such as defects or malformations, will occur. Although DLC1 is generally considered to affect cell motility and focal adhesion via the RhoGap domain and focal adhesion targeting region, respectively, the SAM domain has also been reported to regulate cell migration. We demonstrated that three private variants near the SAM domain could reduce the inhibitory effect of wildtype DLC1, suggesting that these mutations might be implicated in regulating the function of the SAM domain. Although DLC1 isoform 2 has been well studied during the past ten years, the functions of DLC1 isoform 1 still need to be characterized. A series of assays were performed to verify whether DLC1 isoform 1 had a function similar to isoform 2. As shown above, all the mutant and wild-type protein had suppression effects on Rho, and similarly regulated the cytoskeleton rearrangement and prevented the formation of stress fiber in the endothelial cells.