Expression Data Extracted from NCBI GEO (Gene Expression Omnibus) Database

In order to provide proof of concept about the findings of the current study and hypothesis, microarray gene expression data from “Post-mortem Alzheimer’s disease brains: Hisayama Study” available at GEO database were used (reference series {"type":"entrez-geo","attrs":{"text":"GSE36980","term_id":"36980"}}GSE36980) [32]. These data have been generated by microarray analysis using the Affymetrix Human Gene 1.0 ST platform from RNA samples (RIN ≧ 6.9) from gray matter of the frontal cortex (n = 15 AD, n = 18 controls), temporal cortex (n = 10 AD, n = 19 controls), and hippocampus (n = 7 AD, n = 10 controls)] of 88 postmortem brains (https://www.ncbi.nlm.nih.gov/sites/GDSbrowser?acc=GDS4758). All AD cases were pathologically diagnosed as AD or AD-like disorder. For this study, gene expression data for the selected genes, studied in the current manuscript, were extracted and used for correlation analysis.

Similarly, another data set from the GEO database ({"type":"entrez-geo","attrs":{"text":"GSE1297","term_id":"1297"}}GSE1297) was also considered to confirm the correlation analysis observed using the previous data set [33]. This data set contains AD patients of different disease severities (incipient, n = 7; moderate, n = 8; severe, n = 7) and control (n = 9) samples, and gene expression microarray was performed from the hippocampal postmortem tissue (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc={"type":"entrez-geo","attrs":{"text":"GSE1297","term_id":"1297"}}GSE1297). Multiple regression analysis was done using OMI/HTRA2 and MMSE score as dependent variable, while all other expression data of AD-related genes were used as independent variables.