Poster #5 - Biological Embedding of Developmental Psychopathology: Comparing Suitability of Psychological Timing and Constructs in Epigenomic Analysis

• In Event: Poster Session 10

Abstract

BACKGROUND: Child development consists of several periods of transitional stress and psychosocial challenge (Healy et al., 2022). Psychopathology during this time is associated with later adverse psychiatric outcomes (Sallis et al., 2019), which may be facilitated on a molecular level by a process called “biological embedding”. Mounting evidence suggests that early life hardships can become "embedded” into a child’s molecular signature through epigenetic modifications such as DNA methylation (DNAm) (Aristizabal et al., 2020). DNAm is a well-characterized and relatively stable alteration that can influence gene expression without changing the DNA sequence (Smith & Meissner, 2013). Associations between DNAm and numerous aspects of psychopathology have been demonstrated cross-sectionally (Rijlaarsdam et al., 2021) and longitudinally (Meijer et al., 2020); however, most studies only assess a single aspect or timepoint of psychopathology. Given the ubiquity and fluidity of psychopathology across developmental periods (Healy et al., 2022), relying on temporally narrow assessments is considered a limitation in understanding mental health (Caspi et al., 2020). This further hinders interdisciplinary epigenomics work, as it remains unclear which aspect or timeframe of psychopathology represents the most appropriate operationalization of psychopathology development when investigating DNAm. We address this limitation by utilizing multiple longitudinal assessments of a broad and valid measure of psychopathology, the General Psychopathology Factor (GPF) (Gomez et al., 2019), to investigate which timing or trajectory of GPF is most salient for epigenetic embedding.

OBJECTIVE: Determine which timepoint (of 7 timepoints between age 8-18) or trajectory measure (of 5 measures of longitudinal change) of GPF best explains age 18 DNAm patterns from buccal swabs. This is an exploratory characterization without directional hypotheses.

APPROACH (Figure 1): Study participants include 170 children from the Wisconsin Study of Families and Work. Mental health was assessed using mother, teacher, and child responses on the MacArthur Assessment Battery for Middle Childhood (MABMC) (Essex et al., 2002). GPF scores were estimated using confirmatory factor analysis, as previously completed with similar cohorts (Sallis et al., 2019) and the MABMC (Wade et al., 2018). Twelve GPF measures were calculated (Table 1). Buccal swabs were collected at age 18 and DNAm was assessed using the Illumina 450K microarray, measuring ~450,000 DNAm sites per participant.

High-dimensional exploratory linear regression models were used to assess relations between DNAm and GPF. Different models were fit for the twelve GPF measures. All models were adjusted for age, sex, cell-type, genetics, and tobacco use, which are known contributors to DNAm (Teschendorff & Relton, 2018). Models are compared according to a set of statistical criteria suitable for high-dimensional modelling, such as the central tendencies of model fit measures (e.g. median AIC) and graphical assessments for model inflation. The GPF measure that best explains age 18 DNAm will be determined by selecting the best-performing model from these criteria.

SIGNIFICANCE: By characterizing the epigenomic relevance of psychopathology measures, this study provides recommendations for utilizing psychological constructs in epigenome-wide association studies and optimizing resources when designing future longitudinal studies. Determining the most salient timing for biological embedding of psychopathology may also highlight potential intervention targets.

Authors

• Lea Separovic, University of British Columbia (UBC)
  Presenting Author

• Sarah M Merrill, University of British Columbia (UBC)

• Chaini Konwar

• Rebecca Asiimwe

• Julia MacIsaac

• Marilyn J Essex

• W Thomas Boyce, University of California - San Francisco

• Anne M Gadermann

• Michael S. Kobor