| DNA methylation(DNAm)is an important epigenetic event involving the reversible addition of a methyl group primarily to cytosines,without changing genomic sequences.DNAm status is closely associated with cancer,and aberrant methylation can be detected in the early stage of cancer,thus DNAm could be important biomarker for tumor diagnosis and prognosis.Co-occurrence and mutual exclusivity(COME)of DNAm refer to two or more genes that tend to be positively or negatively correlated among different samples.COME of gene mutations in pan-cancer have been well studied,but little is known about the COME of gene methylation in pan-cancer.Here,we first identified some potential biomarkers and build models for the diagnosis and prognosis of pan-cancer.The potential seven diagnostic biomarkers are associated with cell division,DNA replication and cell cycle,the average AUC(Area Under Curve)of the CpG-based diagnostic model can reach above 0.95 when there are more than two CpG sites included in the model.The tumor specific classifier can not only identify different cancer types,but also predict the origin of tumor with metastases or with unknown primary origin.The CpG-based prognostic model built in several cancer types can be used to predict the outcome of patients and classify every patient into high-risk and low-risk group.Then,we systematically explored the COME of gene methylation in diverse human cancer.We found that the genes in functional epigenetic modules identified in the context of COME network of six cancer types were mainly from the zinc finger gene family,and most of those genes were epigenetically repressed.In addition,the frequently occurred COME events can be used to classify the patients into several subtypes with significant different clinical outcomes in six cancer types as well as pan-cancer.Finally,we combined the results of our previous two projects with other related resources to develop a database named DNMIVD,on which researchers can build models for cancer diagnosis and prognosis based on DNA methylation profile as well as to search,visualize and download the resources of CpGs and genes from various aspects.Taken together,our work may facilitate the diagnosis and prognosis of pan-cancer and provoid valuable information for DNA methylation research in cancer. |
