DNA methylation profiling in thyroid cancer in Chinese population

Around 20-30% of Thyroid cancers are difficult to accurately diagnose with fine needle aspiration (FNA) due to their shared features with benign nodules. Although there are molecular tools to assist the diagnosis, such as the Afirma Gene Expression Classifier and ThyroSeqv3, which are the most widely used in the United States, however, these tests all show low positive predictive value and would lead to unnecessary removal of the thyroid in non-cancer patients. In addition, the prediction model of these tests is trained mainly based on the American population. Therefore, it is required to develop a highly accurate test for thyroid nodule diagnostic which could not only increase the PPV, and reduce the unnecessary surgeries, but also is based on Chinese population datasets. To overcome the potential overfitting issues of the prediction model, here we provide a function-guided DNA methylation biomarker selection method through multi-omics datasets. To develop a novel DNA methylation based diagnostic test for thyroid cancer, we characterized the genome-wide DNA methylation pattern in 49 malignant and benign fresh thyroid tissues by using Reduced Representation Bisulfite Sequencing (RRBS) analysis on a single nucleotide level. And we found a group of differentiated DNA methylation sites that could be used to distinguish thyroid cancer and benign tissues. We characterized the genome-wide DNA methylation pattern in 49 malignant and benign fresh thyroid tissues by using Reduced Representation Bisulfite Sequencing (RRBS) analysis on a single nucleotide level.