Schematic of the original experimental design (top) and the current study (bottom).

Percent sum of squares explained by model terms and P-values using all three accessions.

Contrasts of a term are marked with indents. Complete output and further results for individual accessions are given in the supplemental data.

Percent sum of squares explained by model terms and P-values using only Ma-0 and Wei-0 accessions.

Contrasts of a term are marked with indents. Complete output and further results for individual accessions are given in the supplemental data.

Phenotype measures in the G7 generation averaged by the ID of the maternal lines from the G6 generation.

P-values for contrasts with P < 0.05 are indicated (based on the analyses done per accession). P-values with asterisk were not significant (P < 0.05) when outliers were removed (see Figure 3).

Phenotypes measured in the G7 generation.

Brown and green colors represent plants from blocks 1–3 and 4–5, respectively (slower growth in blocks 4– 5). Dots with black strokes were identified as outliers. The number of plants is the minimal number of plants. It varies by 2–3 plants per accession and treatment as some measures were missing. P-values for contrasts with P < 0.05 are indicated (based on the analyses done per accession). P-values with asterisk were not significant (P < 0.05) when outliers were removed.

Gene Ontology enrichment among genes mapped by DMCs.

“Found” indicates the number of different genes with the term that were mapped by DMCs. “Expected” refers to the number of genes that were expected to be mapped by DMCs if DMCs were randomly distributed.

a) Location of the differentially methylated cytosines along the genome. b) Epigenetic variation within different lines (selection replicates), i.e., distances to each centroid. Only lines with at least four individuals are shown. No overall and no pairwise P-value was smaller than 0.05.

a) Genetic distances between the 190 RNA-seq samples from Lin and colleagues (26). The two genetic clusters diverging from the rest of their accession group are highlighted. b) Within accession-label genetic distances, drawn to scale. Maximal distance given below the label, colored according to treatment and generation. The two treatments that have been assigned the wrong accessions are highlighted. c) Pairwise genetic distances within lines (plants of the same lineage, i.e., single-seed descendants), within accessions, and between accessions using either the original accession labels from Lin and colleagues (left, 26) or the accessions assigned due to genetic similarity to the data available at 1001genomes.org (right). The red line demarcates the distance equal to the average of within-line distances plus three standard deviations.