The Use of the Gene Expression Omnibus (GEO) Database to Study Gene Expression and Methylation Patterns in Cases of Preterm Birth and Associated Diseases

Objective

Analysis of gene expression and methylation patterns in preterm births (PTB) and associated diseases from microarray data using the Gene Expression Omnibus (GEO) database. We hypothesise that reported upregulation/downregulation of expression patterns is associated with cases of PTB.

Design

This is a retrospective study done to observe any reported significant changes in gene expression or methylation patterns between PTB, associated diseases, and full-term births (FTB).

Methods

Eight comparative studies on PTB were analysed where cases of FTB were used as controls (GSE178609, GSE48125, GSE197795, GSE190215, GSE150167, GSE166956, GSE188949, GSE225881). Upon accessing the GEO database, cases (PTB) and controls (FTB) were highlighted, and analysis of the data performed. The data was then filtered for statistically significant changes with an adjusted p-value of <0.05 and log fold changes (logFC) ≥1 or ≤-1. Remaining genes after filtering were studied for function and potential relevance to causes of PTB and associated diseases.

Results

Of the 8 datasets, 3 displayed significant changes in expression patterns after application of the filtering criteria. Firstly, the study analysing profiles of umbilical cord exosomes in PTB vs FTB with bronchopulmonary dysplasia (BPD) (GSE190215) showed downregulation in expression of genes in PTB linked to lung function including ciliary function (CFAP61), respiratory insufficiency (CLCN6), and thoracic dysplasia (MORN1). Analysis of the amniotic fluid cell-free transcriptome in spontaneous PTB (GSE166956) showed upregulation of various genes associated with inflammatory processes including IL1A, JAML, SRGN, NFKB, and NAMPT. Finally, the dataset comparing methylation profiles in PTB cases with and without BPD (GSE188949) displayed hypomethylation in cases with no BPD in genes ACP5, linked to lung fibrosis, and STK11IP, associated with lung carcinoma. Hypermethylation of CSTA, identified in lung cancer, was also reported.

Conclusion

The utilization of GEO to analyse gene expression and methylation patterns in cases of PTB and associated diseases has linked various significant genes in potentially impacting such cases. Despite this work being preliminary, it displays a promising avenue of research and through further analysis of multiple datasets can allow for further understanding of the genetic aetiology of PTB. Future work could be done through further understanding of the relevant genes’ function and their pathways and strengthening the validity of the expression patterns in promoting PTB by determining any overlap between the different datasets submitted.