Fig 1.
Flow chart of GEO dataset selection.
The initial search yielded 59 studies. After removing two duplicates, 57 studies were screened for relevance based on the inclusion criteria, which included the use of human clinical samples (excluding human cells/stem cells/iPSCs). Further screening of the remaining 34 studies involved assessing biological matrix, normotensive pregnancies, PE, data sufficiency, absence of other disease states, and the use of the Illumina platform. Five studies met the criteria for deeper screening, of which two were selected for the final analysis.
Fig 2.
(A) Average Estimated Cell Type Distribution and (B) Immune cell proportion between normotensive (grey, control) and preeclamptic (white) pregnancies.
Houseman (2012) was used to deconvolute GSE37722 (N = 14 control, N = 14 preeclamptic pregnancies) to identify changes in the proportion of monocytes, granulocytes, natural killer cells, B-cells, CD4 + T cells and CD8 + T cells. All data is presented as mean. A Kolmogorov-Smirnoff test was used to assess differences in the overall distributions, while a Wilcoxon test was conducted for each cell type. A p-value of ≤ 0.05 was considered statistically significant.
Fig 3.
Average Estimated Cell Type Distribution across gestation for (A) GSE37722 and (B) GSE192918.
Houseman (2012) was used to deconvolute both GEO sets to identify changes in the proportion of monocytes, granulocytes, natural killer cells, B-cells, CD4 + T cells and CD8 + T cells across ‘first trimester, ‘second trimester’, ‘third trimester’ or ‘at delivery’ and ‘postpartum’. While GSE37722 included data for normotensive pregnancy stages, the exact timings of these stages were not explicitly defined in the GEO database. Differences in distribution across gestational were assessed using an Anderson-Darling all pairs comparison test. A p-value of ≤ 0.05 was considered statistically significant.
Fig 4.
Immune cell composition across gestational age within GSE37722.
Houseman (2012) was used to deconvolute both GEO sets to identify changes in the proportion of monocytes, granulocytes, natural killer cells, B-cells, CD4 + T cells and CD8 + T cells across ‘first trimester, ‘second trimester’, ‘at delivery’ and ‘postpartum’. While GSE37722 included data for normotensive pregnancy stages, the exact timings of these stages were not explicitly defined in the GEO database. All data is presented as mean. A linear model was conducted for each cell type. An FDR-adjusted p-value of ≤ 0.05 was considered statistically significant.
Fig 5.
Immune cell composition across gestational age within GSE192918.
Houseman (2012) was used to deconvolute both GEO sets to identify changes in the proportion of monocytes, granulocytes, natural killer cells, B-cells, CD4 + T cells and CD8 + T cells across ‘first trimester, ‘second trimester’, ‘third trimester’ and ‘postpartum’. GSE192918 defined these stages as follows: early-pregnancy (10-14 weeks), mid-pregnancy (24-28 weeks), at delivery (38-40 weeks), and postpartum (10 month post-delivery). All data are presented is mean ± SEM. A linear model was conducted for each cell type. An FDR-adjusted p-value of ≤ 0.05 was considered statistically significant.