A meta-analysis of epigenome-wide association studies on pregnancy vitamin B12 concentrations and offspring DNA methylation.

"the dataset(s) supporting the conclusions of this article is available in the [zenodo repository]. data availability statement analysis plan and r code for cohort-specific analyses and meta-analyses are available via https://github com/giuliettamonasso/pace-b12-meta-analysis-of-ewas . cohort-specific characteristics are shown in supplementary data 1-3 .; across cohorts with these data the correlations of maternal vitamin b12 with folate ( r < 0 15) and homocysteine ( r < -0 26) were low (supplementary data 4) .; supplementary data 5 shows the lambdas of all cohort-specific analyses and meta-analyses.; in a look-up of the 109 prioritized cpgs from the maternal meta-analysis19 (17 6%) of 108 available cpgs were also associated with newborn vitamin b12 concentrations at bonferroni-significance ( p -value <0 05/108 tests i e p -value <4 63 x 10 -4 ; supplementary data 6 ) and all had the same direction of effect.; the seven newborn prioritized cpgs were not associated with maternal vitamin b12 concentrations at bonferroni significance ( p -value <0 05/7 tests; i e p -value <0 007) although one had an uncorrected p -value <0 05 and all had the same direction of effect ( supplementary data 7 ).; the direction of the effect was consistent between alspac and genr for 5/7 prioritized cpgs ( supplementary data 8 ) [ ].; early-pregnancy associations were largely consistent with anytime associations with pearson's correlation coefficients for the effect estimates r = 0 78 (epigenome-wide) and r = 0 99 (prioritized cpgs) ( supplementary data 9 ).; of these two cpgs were differentially methylated in relation to maternal pregnancy circulating vitamin b12 concentrations ( p -value <0 05/103 tests i e p -value <4 85 x 10 -4 ; supplementary data 10 ) and 56/103 (54 4%) cpgs had consistent direction of effect.; in late childhood (ages9-10; n = 482) 7 (6 5%) cpgs were still associated with maternal early-pregnancy circulating vitamin b12 concentrations ( table 2 and supplementary data 11 ).; in blood dna methylation data sampled in late childhood ( n = 321 genr) and adolescence (age 17; n = 83 alspac) 4 (57 1%) and 0 cpgs respectively were still associated with newborn vitamin b12 concentrations ( supplementary data 12 ).; in the maternal meta-analysis ( n = 2397 all cohorts) 89/109 (81 7%) findings remained significant at epigenome-wide level ( p fdr <0 05 supplementary data 9 ).; although all seven findings had p fdr >0 05 pearson's correlation for the effect estimates between the primary and folate-adjusted models was high ( r = 0 95: epigenome-wide; r = 0 99: prioritized cpgs supplementary data 8 ).; in the maternal meta-analysis only the top hit cg25327343 remained fdr-significant at epigenome-wide level ( n = 2020; meta-analysis without inma; uncorrected p -value = 2 31 x 10 -8 ; supplementary data 9 ).; although all seven findings had p fdr >0 05 pearson's correlation coefficient for effect estimates between the primary and homocysteine-adjusted models was high ( r = 0 90 epigenome-wide; r = 0 99: prioritized cpgs supplementary data 8 ).; also 10/15 had p fdr <0 05 in our folate-adjusted meta-analysis ( supplementary data 9 ) [ ].; of the prioritized cpgs in the maternal and newborn meta-analysis4/109 ( p -value <0 05/109 tests; i e p -value <4 59 x 10 -4 ) and 1/7 cpgs ( p -value <0 05/7 tests; i e p -value <0 007) respectively were also differentially methylated in relation to birth weight with a similar direction of effect ( tables2-3 and supplementary data 13-14 ) [ ].; of the prioritized cpgs in the maternal meta-analysis1/109 cpgs (cg27181142) was also differentially methylated in relation to gestational age with a similar direction of effect ( tables2-3 and supplementary data 13-14 ) [ ].; none of the prioritized cpgs in the maternal meta-analysis were differentially methylated in relation to childhood overall cognitive skills and childhood nonverbal iq ( supplementary data 13 ) [ ].; of these 1/5 and 2/5 cpgs respectively were differentially methylated in relation to these traits with a similar direction of effect ( supplementary data 14 ).; follow-up analyses of the identified cpg sites the 109 and 7 prioritized cpgs from the maternal and newborn meta-analysis respectively showed little evidence for functional enrichment of go terms (smallest p -value = 9 8 x 10 -4 ) or kegg terms (smallest p -value = 4 9 x 10 -3 ) terms ( supplementary data 15-19 ).; 500kb of any of 20/109 (18 3%) prioritized cpgs ( tables1-2 ; supplementary data 19 )."

"data availability statement analysis plan and r code for cohort-specific analyses and meta-analyses are available via https://github com/giuliettamonasso/pace-b12-meta-analysis-of-ewas ."

"Disclosure statement No potential conflict of interest was reported by the authors."

"For all studies, acknowledgements can be found in Supplementary Information: Acknowledgements. For all studies, funding statements can be found in Supplementary Information: Funding."