Investigators at Vanderbilt University Medical Center and the University of Calgary have devised a new approach for conducting gene-based analyses for cancer susceptibility that outperforms existing models.
Their research, published online Nov. 19 in Nature Communications, tested a new model for transcriptome-wide association studies (TWAS), a gene-based method for investigating susceptibility to cancer and other complex diseases. The researchers integrated transcription factors, which are the proteins involved in the process of gene regulation, into a new model called sTF-TWAS. The transcription factors they chose have been previously identified to be associated with cancer susceptibility.
The researchers then applied this model to data from large genome-wide association studies that identify putative susceptibility genes associated with phenotypic variations in disease manifestation. They focused on three common cancers: breast cancer, prostate cancer and lung cancer.
The new approach detected more genes than those from existing TWAS models. The investigators identified a total of 354 cancer susceptibility genes, including 234 that had been unreported.
“With integration of prior information of transcription factor occupancies into TWAS, our approach can significantly improve cancer susceptibility gene discovery. Our approach can also be easily expanded to genetically predicted DNA methylation and proteins (i.e., proteome-wide association study) in future genetic studies,” said Xingyi Guo, PhD, associate professor of Medicine in the Division of Epidemiology at VUMC, who conceived and designed the study with Quan Long, PhD, associate professor of Biochemistry and Molecular Biology at the University of Calgary.
The team of researchers also tested whether the sTF-TWAS model could be utilized for non-cancer diseases, conducting analyses for the brain disorders, schizophrenia, Alzheimer’s and autism spectrum disorder. They found that sTF-TWAS identified more susceptibility genes for each disorder than S-PrediXcan, another gene-mapping model used to identify disease associations.
“We also show the generalizability of the sTF-TWAS on non-cancer diseases. Our sTF-TWAS approach will not only strengthen the detection of novel disease genes, but also advance the understanding of transcription factor-based transcriptional networks underlying genetic susceptibility to human diseases,” Guo said.
The sTF-TWAS approach also outperformed current existing approaches such as S-PrediXcan in identifying cancer susceptibility genes.
A limitation of the study is that the data registries utilized for breast cancer, prostate cancer and lung cancer were compiled from European descendants.
The research received support from U.S. National Institutes of Health grants (R37 CA227130 to Xingyi Guo and R01 CA235553 to Wei Zheng) and a Canadian grant from the New Frontiers in Research Fund.
Other Vanderbilt authors on the study are Wanqing Wen, MD, MPH; Xiao-Ou Shu, MD, PhD, MPH, and Wei Zheng, MD, PhD, MPH.
