The invention discloses a biomarker selecting method based on meta-analysis, and a
system. The method comprises the following steps of selecting original
sequencing data; performing mapping analysis on the original
sequencing data by means of an FANSe
algorithm, obtaining
gene quantification information and setting a
gene original group; calculating an importance sequence of the
gene in the original group according to a GWGS
algorithm, integrating the importance of each group of genes according to a GWRS
algorithm, obtaining an integrated gene importance arrangement
list, sequencing the genes according to the importance from highest to lowest; performing
data mining by means of a Wrapper
Feature Selection model based on the SVM, differentiating a data
sample type, and screening the biomarker from the gene with high importance. According to the method, multi-center original
sequencing data are integrated according to the characteristic of the sequencing data, and
system differences in platform, sample and experiment designing are settled; an integrating analysis algorithm with high robustness is used for performing deep
data mining, thereby mining common, specific and key biomacromolecules.