Meta-analysis of placental gene expression
Six PE placenta expression studies from Gene Expression Omnibus (GEO)
datasets were combined and subjected to multiplex analysis with the
method as we previously developed . We calculated the meta-fold change
of each gene across all studies. For gene expression measurements, this
corresponds to combining fold-changes across studies to identify a
meta-fold-change that is an amalgamation of the constituent studies. The
meta p-values were obtained by Fisher’s method. Significant genes were
selected if they were measured in five or more studies and the meta
effect p value was less than 4.5×10-5. This effort
identified Lep as the leading protein biomarker candidate. We also
compared the expression of Lep transcript between non-PE and PE women in
individual datasets.