Analysis Of Mrna Expression Profiles Data To Identify And Verify Key Genes In SLE

Objective: We use the Gene Expression Omnibus(GEO)to search for DNA microarrays that related to Systemic Lupus Erythematosus(SLE),download the chip data,conduct meta-analysis,summarize the SLE gene marker collection,and verify them.Therefore,it can provide reference and theoretical basis for determining the key genes of SLE and disease prevention,early diagnosis and drug target treatment.Methods: DNA microarray data were downloaded from the NCBI(National Center for Biotechnology Information)GEO(Gene Expression Omnibus)database and Pubmed database,and 15 DNA microarrays for the SLE phenotype were identified and then divided into a training set(13 DNA microarrays)and a test set(2 DNA microarrays),including 1,869 and 1,171 samples(see table S1)respectively.Key genomics were extracted from 13 DNA microarrays in the training group,we performed a meta-analysis using EXALT(Expression Analysis Tool)software,and then retrospectively and prospectively verified it by using samples from two DNA microarrays in the test group.Compare the expected SLE predictions in the test group data set with the actual clinical diagnosis.The primary predicted end point is the SLE diagnosis or SLE disease activity(DA)of the validation cohort,the positive predictive value(PPV),the negative prediction value(NPV),sensitivity and specificity that were generated by statistics and were evaluated for predictive performance.Receiver operating characteristic(ROC)curve analysis was performed using statistical software R(Version 3.3.3)to determine the sensitivity and specificity of the SLEmetaSig100 prediction and to determine the area under curvev(AUC),and then to evaluate the ability of the selected feature gene sets to predict SLE.Results: From 13 training data sets on SLE gene-expression studies,we identified a SLE meta-signature(SLEmetaSig100)containing 100 concordant genes that are involved in DNA sensors and the IFN signaling pathway.We rigorously examined SLEmetaSig100 with both retrospective and prospective validation in two independent data sets.We found that in the SLE dominant sample groups,there was only one(out of 72 total)normal sample in GSE65391 and none(out of 20 total)in GSE11909,while remaining normal samples(71 in GSE65391 and 20 in GSE11909)were grouped in the normal sample groups.The centroid model can further prospectively apply to individual patients with high PPV(97%-99%),specificity(84%-85%),and sensitivity(60%-84%).The ROC results are comparable displaying areas under the curve(AUC)of 0.89(GSE65391)and 0.85(GSE11909),respectively,Conclusions: SLEmetaSig100 plays an important role in the occurrence of SLE.It may be a key gene collection of SLE.It has strong predictive value for SLE and can promote the current clinical diagnosis of SLE,and also provide personalized immune monitoring of SLE disease activities.Therefore,it provides a theoretical basis for determining the key genes of SLE and disease prevention,early diagnosis and drug target treatment.