A novel hypothesis-generating approach for detecting phenotypic associations using epigenetic data

Abstract

AbstractBackgroundHypotheses about what phenotypes to include in causal analyses (that in turn can have clinical and policy implications) can be guided by hypothesis-free approaches, leveraging the epigenome for example. Materials & methods: Minimally adjusted epigenome-wide association studies (EWAS) using ALSPAC data were performed for example conditions, dysmenorrhea and heavy menstrual bleeding (HMB). Differentially methylated CpGs were searched in the EWAS Catalog and associated traits identified. Traits were compared between those with and without the example conditions in ALSPAC. Results: Seven CpG sites were associated with dysmenorrhea and two with HMB. Smoking and adverse childhood experience score were associated with both conditions in the hypothesis-testing phase. Conclusion: Hypothesis-generating EWAS can help identify associations for future analyses.Plain language summaryTo make a positive impact on policy and clinical practice, it is important that epidemiologists, those who study population health, can identify characteristics that might increase the risk of medical conditions. However, it can be difficult to know which associations should be investigated and decisions can often be biased by pre-formed opinions about what is relevant. In this study, we wanted to look for potential risk factors for dysmenorrhea (painful periods) and heavy menstrual bleeding (HMB) using a hypothesis-free approach (in other words, minimal adjustment for potential confounders), leveraging epigenetic data from a sub-sample of the Avon Longitudinal Study of Parents and Children (ALSPAC) and generating hypotheses about associations, then testing these hypotheses in the wider ALSPAC cohort. This meant looking for differentially methylated CpGs between those with and without the conditions of interest using an epigenome-wide association study (EWAS), seeing which phenotypes were associated with the CpGs in the EWAS Catalog, and testing these hypotheses in the ALSPAC cohort using measurements of each phenotype. For dysmenorrhea, we found seven differentially methylated CpGs and for HMB, we found two. These CpGs were associated with several phenotypes, which we could proxy in the wider ALSPAC cohort, creating hypotheses we tested using regression analyses. In the hypothesis-testing phase, we found that smoking and adverse childhood experience score were associated with dysmenorrhea and HMB. With this under-utilised approach, we can identify phenotypes that may be risk factors for under-studied conditions, that can be explored in other cohorts using analyses that can assess causality.Tweetable abstractLeveraging EWAS data can help identify novel potential risk factors for understudied conditions such as dysmenorrhea and heavy menstrual bleeding for future examination in causally motivated analyses: a proof-of-concept study in the Children of the 90’s cohort (ALSPAC)