Partial Reproducibility and Pleiotropic Epigenetic QTLs in Arabidopsis Recombinant Inbred Lines

Unlike conventional genetic polymorphisms, many induced epigenetic polymorphisms can be reset across generations, raising uncertainty about how consistently the same epigenetic loci and associated phenotypic effects can be recovered among independent studies. To address this problem, we designed our study specifically to maximize a lineage-matched, environmentally aligned cross-study comparability with the foundational work of Cortijo et al. by using seeds derived from the same epiRIL generation and grown under broadly similar environmental conditions. We mapped flowering time, as well as five non-flowering traits that had not previously been mapped in this epiRIL population: rosette diameter, basal branch number, lateral branch number, fruit number, and rosette leaf number. We detected significant epiQTLs for all traits except lateral branch number. We also reproduced a facsimile of the original computational pipeline. This design provides an approximate upper bound on expected reproducibility before additional generations of methylation resetting or divergence among seed stocks could substantially influence the results. We recovered two of Cortijo et al.'s previously reported epiQTLs, failed to recover another, and found a different one, with power analyses suggesting that the discrepancies may reflect statistical power. Within overlapping epiQTL intervals across the five traits that had significant intervals, we identified 69 candidate genes exhibiting gene body methylation. Overall, our results reproduced some but not all previous epiQTL signals when comparisons were made using closely matched source material under similar conditions.