Research On Precise Extension Algorithm Of Methylation Chip Data And Implementation Ofprediction Platform

DNA methylation plays an important role in regulating gene expression and is an important subject in epigenetics.The methods for detecting DNA methylation levels generally include sequencing technology and chip detection technology.The sequencing technology can obtain the methylation level of human whole genome CpG sites,but the cost is high and the means are complicated;the chip detection technology is relatively inexpensive and capable of Obtaining genome-wide DNA methylation levels is currently the primary method for detecting DNA methylation levels.The downside of the mainstream 450 K methylation chip is that it covers only 2% of the total number of human genome-wide CpG sites.Therefore,the use of computational methods to expand the 450 K methylation chip data to obtain more methylation levels of CpG sites has become a research hotspot.The current 450 K methylation chip data expansion method belongs to the development of a generalized extended model,which cannot quantify the prediction accuracy of a specific site.To this end,this paper designed a method to accurately expand the 450 K methylation chip data,build an extended model for a single CpG site and develop an online extension platform.In addition,the extended model was applied to the methylation level of circular RNA,and the association between methylation of circular RNA and cancer was analyzed.The specific work of this paper is as follows:1.Proposed a precise extension algorithm of 450 K methylation chip data based on methylation level similarity and DNA sequence composition similarity measure between CpG sites.An extended model is established for a single CpG site,and data training is performed by WGBS.The model was tested to obtain the parameters,errors and prediction accuracy of the model.Finally,a precise expansion model for the methylation level of a single site was obtained.The correlation coefficient between the prediction result and the WGBS detection result reached 0.93,and the performance indexes were excellent.Compared with the existing methylation level expansion model,this model can quantitatively evaluate the prediction effect of specific sites.2.Since the methylation level of only a few sites in the CpG site covered by the circular RNA is detected,it is impossible to systematically analyze the methylation pattern in the circular RNA at this stage.The precise extension model is applied to the circular RNA,and the methylation profile of the circular RNA is systematically analyzed based on the extended circular methylation data.Combined with the multi-omics analysis method,some cancers are closely related.The circular RNA provides a basis for biologists to select experimental objects from massive data.3.Designed and developed a DNA methylation online prediction platform based on Django framework.The platform has the characteristics of high response speed,high stability and high concurrency,and is easy to access and simple to use.It provides relevant scientific research workers at home and abroad.A scientific tool that can predict the methylation level of a specific CpG site online,enriches the means of obtaining DNA methylation levels,and saves the overhead of methylation detection experiments.