To quantify the amount of variation in DNA methylation explained by genomic context, we considered the correlation between genomic context and principal components (PCs) of methylation levels across all 100 samples (Figure 4). We found that many of the features derived from a CpG site’s genomic context appear to be correlated with the first principal component (PC1). The methylation status of upstream and downstream neighboring CpG sites and a co-localized DNAse I hypersensitive (DHS) site are the most highly correlated features, with Pearson’s correlation r=[0.58,0.59] (P<2.2?10 ?16 ). Ten genomic features have correlation r>0.5 (P<2.2?10 ?16 ) with PC1, including co-localized active TFBSs ELF1 (ETS-related transcription factor 1), MAZ (Myc-associated zinc finger protein), MXI1 (MAX-interacting protein 1) and RUNX3 (Runt-related transcription factor 3), and co-localized histone modification trimethylation of histone H3 at lysine 4 (H3K4me3), suggesting that they may be useful in predicting DNA methylation status (Additional file 1: Figure S3). 67,P<2.2?10 ?16 ) [53,54].

Relationship matrix out of forecast provides with very first 10 Personal computers out of methylation profile. The fresh new x-axis represents among the many 122 has actually; brand new y-axis represents Personal computers step 1 courtesy 10. Tone match Pearson’s correlation, because the found throughout the legend. Pc, dominant part.

Digital methylation status prediction

These observations about patterns of DNA methylation suggest that correlation in DNA methylation is local and dependent on genomic context. Using prediction features, including neighboring CpG site profily buddygays methylation levels and features characterizing genomic context, we built a classifier to predict binary DNA methylation status. Status, which we denote using ? _i,j ? <0,1>for i ? <1,...,n> samples and j ? <1,...,p> CpG sites, indicates no methylation (0) or complete methylation (1) at CpG site j in sample i. We computed the status of each site from the ? _i,j variables: \(\tau _ = \mathbb <1>[\beta _ > 0.5]\) . For each sample, there were 378,677 CpG sites with neighboring CpG sites on the same chromosome, which we used in these analyses.

Therefore, prediction out of DNA methylation standing depending merely towards methylation levels at the surrounding CpG internet may well not work well, particularly in sparsely assayed areas of the newest genome

The newest 124 possess that we used for DNA methylation standing anticipate get into four some other kinds (look for A lot more file step one: Desk S2 to have an entire number). For every single CpG website, we include the following function establishes:

neighbors: genomic ranges, digital methylation condition ? and you will accounts ? of one upstream and you to downstream surrounding CpG web site (CpG web sites assayed into the range and you will adjacent regarding genome)

genomic status: binary philosophy indicating co-localization of your CpG web site having DNA sequence annotations, and additionally promoters, gene body, intergenic area, CGIs, CGI shores and you may cabinets, and you can close SNPs

DNA sequence qualities: continued values symbolizing neighborhood recombination speed out-of HapMap , GC content off ENCODE , integrated haplotype results (iHSs) , and you may genomic evolutionary rates profiling (GERP) calls

cis-regulating issue: binary opinions exhibiting CpG website co-localization which have cis-regulating factors (CREs), and DHS web sites, 79 specific TFBSs, 10 histone modification scratching and you can 15 chromatin states, most of the assayed regarding the GM12878 mobile line, the fresh closest match so you’re able to entire bloodstream

We used a RF classifier, which is an ensemble classifier that builds a collection of bagged decision trees and combines the predictions across all of the trees to produce a single prediction. The output from the RF classifier is the proportion of trees in the fitted forest that classify the test sample as a 1, \(\hat <\beta>_\in [0,1]\) for i=<1,...,n> samples and j=<1,...,p> CpG sites assayed. We thresholded this output to predict the binary methylation status of each CpG site, \(\hat <\tau>_ \in \<0,1\>\) , using a cutoff of 0.5. We quantified the generalization error for each feature set using a modified version of repeated random subsampling (see Materials and methods). In particular, we randomly selected 10,000 CpG sites genome-wide for the training set, and we tested the fitted classifier on all held-out sites in the same sample. We repeated this ten times. We quantified prediction accuracy, specificity, sensitivity (recall), precision (1? false discovery rate), area under the receiver operating characteristic (ROC) curve (AUC), and area under the precision–recall curve (AUPR) to evaluate our predictions (see Materials and methods).