Epigenome target exome sequencing and cis-meQTL mapping in a giga-genome conifer

A fundamental question has been whether variation in plant DNA methylation is chiefly driven by environmental cues or under genetic control, and therefore heritable and accessible to selection, with important consequences for epigenetic breeding and for understanding heritable phenotypic plasticity. In this work, we have addressed this question by cis-meQTL mapping across all three cytosine contexts in 602 Picea abies individuals, pairing exome capture sequencing (ECS) and targeted bisulfite sequencing (TBS) across the same genomic regions in a pedigree-structured breeding cohort (209 individuals, 17 full-sib families) and natural populations (393 individuals, 25 natural stands). Retaining only strongly significant associations yielded 6,954 and 18,480 SNP–methylation pairs in the breeding and natural cohorts, respectively (FDR < 1×10⁻¹⁰), indicating that a substantial fraction of the methylome is genetically determined and classically heritable. Genomic-epigenomic concordance was significant in the pedigree-structured cohort while near-absent in wild stands, revealing that genetic control of methylation is detectable when relatedness is structured but masked by environmental and stochastic variance in natural populations. Robust meQTL SNPs were almost entirely cohort-specific (< 0.5% shared between breeding and natural cohorts across all three cytosine contexts), indicating that the genetic architecture of cis-methylation control operates through largely non-overlapping loci in structured pedigrees versus geographically distributed populations. CHG methylation showed the strongest family-level differentiation, consistent with its self-reinforcing inheritance through H3K9me2–CMT3 coupling. The most pleiotropic meQTL source loci clustered at NBS-LRR immune receptor gene arrays and a cis-regulated histidine kinase locus in the natural cohort controlling 68 downstream methylation sites. Strikingly, methyl-CpG binding domain (MBD) reader proteins were recurrently identified among the most significant source genes. Moreover, RNA-dependent RNA polymerases and an Argonaute protein emerged as CHH-context targets, implicating feedback regulation of the RdDM silencing pathway. By using two complementary association frameworks, we identified thousands of robust meQTLs and generated the first high-confidence map of genetically determined methylation loci in a large-genome species. These findings demonstrate that plant DNA methylation variation is substantially genetically encoded and provide a high-confidence atlas of heritable methylation loci as a foundation for incorporating epigenomic markers into genomic selection pipelines for long-lived trees.