Figure 3

DNA methylation levels in myocytes at the distal D4Z4 repeat on the contracted 4qA chromosome correlate with disease. BSS analysis of the distal pathogenic D4Z4 RU in family cohorts of myogenic cells derived from biceps (cohorts 03, 07, 09, 12, 16, 17, 19, and 21) of FSHD1-affected (left column) and unaffected (right column) subjects. Overall, 56 predicted CpGs (each represented as a column, numbered 1-56, left to right) arranged linearly on a chromosome were assayed for the 4qA BSS assay, or 30 predicted CpGs (columns numbered 1-30, left to right) for the 4qA-L BSS assay, as indicated in the left margin. Each independent chromosome assayed is represented by a row with each CpG represented by a box (red boxes indicating methylation, blue boxes indicating lack of methylation, and empty boxes indicating lack of a CpG detected at that site). Importantly, on average >99% of the predicted CpGs were identified in the sequences analyzed for each sample, and each total sequence had >98% identity to the reference sequence, indicating that the amplified BSS products are specific to 4qA and there are very few polymorphisms. The haplotypes, associated EcoRI/BlnI fragment sizes, and calculated D4Z4 RUs of the shortest FSHD-permissive allele are listed after sample names taken from Additional file 1: Table S1; symbols ^ and ^^ are described there. Numbers in the right margin indicate estimated percent methylation for each of two alleles using a beta-binomial mixture model (allele 1 in red, allele 2 in green), and using a mono-allelic model (orange). The color bar in the right margin indicates confidence in assignment of each sequence to each allele: pure red for 100% posterior probability for allele 1; pure green for 100% posterior probability for allele 2; blended colors for intermediate values (color scale in the lower right).