Additional Supporting Information may be found in the online version of this article.

bdra23126-sup-0001-suppinfo.tif103KFigure S1. A schematic outline of the Selective Enrichment of Methylated DNA (SEMD) protocol used for microarray analysis. Genomic DNA was isolated from GD 12–14 secondary palates and sonicated to generate fragments in the 300 bp – 1 kbp range. Sonicated DNA was treated with MBD2b, a methyl-CpG binding protein (Active Motif, Inc., CA) and the bound DNA fragments were released and amplified. Sonicated DNA not subjected to SEMD was similarly amplified (input DNA). Input DNA, labeled with Cy3, and SEMD DNA, labeled with Cy5, were hybridized to NimbleGen 2.1M mouse promoter microarrays (Roche NimbleGen, WI). Nine microarrays representing 3 biological replicates from each of GDs 12, 13 and 14, were probed. Data accruing from such hybridization were analyzed as detailed in Methods.
bdra23126-sup-0002-suppinfo.tif110KFigure S2. Boxplots. Figure S1 presents boxplots of all nine arrays for each channel and for the log2 red/green intensity ratios, after T-quantile normalization. Each box corresponds to a single array. The arrays are well-aligned in each channel, as expected from T-quantile normalization.
bdra23126-sup-0003-suppinfo.tif435KFigure S3. MA plots. Figure S2 shows the ‘MA’ plot for each array, after T-quantile. Here, M and A are defined as: M = log2(I1) - log2(I2); A = 1/2 (log2(I1)+log2(I2)), where I1 is the log2 red/green intensity ratio of the array studied and I2 is the log2 red/green intensity ratio of a “pseudo”-array, which consists of the median values of all the arrays. The mass of the distribution is expected to be concentrated along the M = 0 axis, and there should be no trend in the mean of M as a function of A. Outlier detection was performed by computing Hoeffding's statistic Da, which tests for the independence of the A and M distributions for each array. Shown are the 4 arrays with the highest value of Da (top row), and the 4 arrays with the lowest value (bottom row). The value of Da is shown in the panel headings. None of the arrays had Da values larger than the threshold for outlier detection (0.15).
bdra23126-sup-0004-suppinfo.docx15KSupporting Information Table S1
bdra23126-sup-0005-suppinfo.xlsx996KSupporting Information Table S2
bdra23126-sup-0006-suppinfo.xlsx13KSupporting Information Table S3
bdra23126-sup-0007-suppinfo.xlsx160KSupporting Information Table S4
bdra23126-sup-0008-suppinfo.xlsx147KSupporting Information Table S5
bdra23126-sup-0009-suppinfo.xlsx160KSupporting Information Table S6
bdra23126-sup-0010-suppinfo.xlsx18KSupporting Information Table S7
bdra23126-sup-0011-suppinfo.xlsx18KSupporting Information Table S8
bdra23126-sup-0012-suppinfo.xlsx19KSupporting Information Table S9
bdra23126-sup-0013-suppinfo.docx16KSupporting Information Table S10

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.