The invention discloses a
methylation vector-based
lung cancer subtype specific
methylation region identification method, a biomarker and application, and relates to the technical field of
sequence analysis, the method comprises the following steps: obtaining and preprocessing
bisulfite sequencing data of a to-be-analyzed sample; sliding a window on the
reference genome for scanning, and coding reads covering all CpG sites in the window into
methylation vectors according to
a site sequence; collecting same window vectors, clustering to form a vector cluster, calculating a group specificity vector proportion and a sample proportion, screening according to rules, combining adjacent window intervals, and outputting subtype specificity methylation vector areas; the first 500 genes are screened from a low-methylation region as biomarkers and are used for
molecular typing and
early detection, and prognosis evaluation and
curative effect monitoring are supported. Through the technical scheme of the invention, the read-level joint methylation configuration is reserved, the coverage fluctuation and heterogeneity interference are reduced, and the stability and discrimination precision of region identification are improved.