Construction of an immune-related ceRNA network to screen for potential diagnostic markers for autism spectrum disorder.

The diagnosis of autism spectrum disorder (ASD) is reliant on evaluation of patients' behavior. We screened the potential diagnostic and therapeutic targets of ASD through bioinformatics analysis. Four ASD-related datasets were downloaded from the Gene Expression Omnibus database. The "limma" package was employed to analyze differentially expressed messenger (m)RNAs, long non-coding (lnc)RNAs, and micro (mi)RNAs between ASD patients and healthy volunteers (HVs). We constructed a competing endogenous-RNA (ceRNA) network. Enrichment analyses of key genes were undertaken using the Gene Ontology database and Kyoto Encyclopedia of Genes and Genomes database. The ImmucellAI database was used to analyze differences in immune-cell infiltration (ICI) in ASD and HV samples. Synthetic analyses of the ceRNA network and ICI was done to obtain a diagnostic model using LASSO regression analysis. Analyses of receiver operating characteristic (ROC) curves were done for model verification. The ceRNA network comprised 49 lncRNAs, 30 miRNAs, and 236 mRNAs. mRNAs were associated with 41 cellular components, 208 biological processes, 39 molecular functions, and 35 regulatory signaling pathways. Significant differences in the abundance of 10 immune-cell species between ASD patients and HVs were noted. Using the ceRNA network and ICI results, we constructed a diagnostic model comprising five immune cell-associated genes: adenosine triphosphate-binding cassette transporter A1 (), DiGeorge syndrome critical region 2 (), glucose-fructose oxidoreductase structural domain gene 1 (), glutaredoxin (), and SEC16 homolog A (). The diagnostic performance of our model was revealed by an area under the ROC curve of 0.923. Model verification was done using the validation dataset and serum samples of patients. , and could be diagnostic biomarkers and therapeutic targets for ASD.