Postmortem anterior cingulate lipid composition in people who died by suicide: exploratory multivariate findings shaped by age and environmental factors

Suicide is a major public health concern, and lipid metabolism has been implicated in its neurobiology. We profiled fatty-acid composition in bilateral pregenual anterior cingulate cortex (ACC) samples obtained at forensic autopsy from 98 individuals (42 who died by suicide, 56 sudden-death controls) using targeted GC-MS and centered log-ratio (CLR) transformation. Analyses included exploratory univariate tests, covariate-adjusted linear mixed models (LMMs; FDR correction), and Bayesian Additive Regression Trees (BART) classification. Univariate tests showed nominal case-control differences in cholesterol and several n-6 PUFAs, but no lipid showed a significant main group effect after covariate-adjusted LMMs with FDR correction. Age and season were more consistent correlates of ACC lipid composition. In the full sample, a complete-case lipids-only BART model achieved CV-BCA = 72.0 ± 5.1% ( n = 81; permutation p < 0.001), but interpretation was limited by different analytic samples across model types. In an adult-only sensitivity analysis (age ≥18 years; n = 92), covariate-adjusted LMMs again showed no FDR-significant group effects, whereas an exploratory mixed BART model (sex, age, PMI, ECFA 20:0, DPAn-6, EPA; n = 75 complete cases) achieved CV-BCA = 75.5 ± 3.7% and nested CV-BCA = 73.5 ± 13.3%. These findings do not support robust single-lipid suicide-associated biomarkers. Rather, they suggest that ACC lipid signals are subtle, multivariate, and strongly shaped by demographic, seasonal, and postmortem factors. The study provides a methodological caution for postmortem lipidomics and requires external validation in clinically characterized cohorts.