EPI-2021-0220 Suppl. Data

Fig 1. Scree plot showing the proportion of the total variation in the individual glucose and insulin values captured by the principal components (PCs).

Fig 2. Histogram showing the distribution of top three PCs derived from maternal glucose and insulin values measured at each time-point (fasting, 1-h and 2-h) during a 75-g OGTT (2^nd trimester).

Fig 4. Plot of loading values showing the position of measured glycemic traits (OGTT) in the two-dimensional space for three components estimated using a PC analysis.

Fig 5. Summary plots of the EWAS of PC1 and cord blood DNA methylation (in M-values) (N = 430).

Fig 6. Summary plots of the EWAS of PC2 (A) and PC3 (B) in association with cord blood DNA methylation (in M-values)

Fig 7. Manhattan plot summarizing results of the EWAS of six individual glycemic traits measured during the OGTT in the second trimester

Table 2. Pairwise Pearson correlations of measured glycemic traits in the OGTT (fasting, 1-h and 2-h post-load insulin and glucose levels) and top three principal components (PCs) derived from these variables, with different birth outcomes.

Table 3. Additional adjustment for maternal BMI in early pregnancy at PC1-associated CpG sites in cord blood previously detected in our main EWAS model using as covariates maternal age, gestational week at the 2^nd visit, gravidity, gestational age at delivery (weeks), child sex, maternal smoking in pregnancy and estimated cell-type distribution (7 cell types from cord blood).

Table 4. Look-ups of previously reported associations of cord blood DNAm with gestational diabetes (GDM) or maternal hemoglobin A1c (HbA1c) in pregnancy, in our EWAS of PC1