Extended Data Figure 3. Mapping m⁶A distribution in the transcriptome of paused ESCs.

a. Strategy for MeRIP-seq in ESCs with cell number-normalization (CNN) using human cell spiking. b. Validation of the CNN strategy for the MeRIP-seq. By mixing different ratios of human cells to ESCs (1, 2 or 4%), we simulated global changes in methylation. Spiking normalization allows capture of these differences, as shown here by MeRIP-qPCR for 3 methylated mRNAs (NeuroD1, Nr5a2, Sox1). Data are mean ± SD, n=5 biological replicates with levels relative to 2% spike in each replicate. c. The specificity of the m⁶A capture was tested by spiking poly(A) RNA from ESCs with exogenous RNAs before performing MeRIP-qPCR. Data are mean ± SD, n=3 biological replicates. P-values (as indicated on figure) by two-tailed unpaired Student’s t-tests. d. Examples of gene track views of MeRIP-seq, for mRNAs of pluripotency factors previously shown to be methylated in ESCs. e. Motif analysis performed with DREME in a 100bp window surrounding MeRIP peak summits identifies several motifs corresponding to the consensus “DRACH” m⁶A motif (where D=A, G or U; H=A, C or U). f. Distribution of differential m⁶A peaks, according to the type of structural element within the transcript. g. Examples of gene track views of MeRIP-seq, for mRNAs with significant hypermethylation (Tial1, Ptrf) or hypomethylation (Cenpt) in pausing of ESCs. h. GSEA of m⁶A changes in paused ESCs relative to control ESCs, using the “hallmarks” collection. No single pathway is significantly enriched based on m⁶A changes (representative pathways are shown). P-values (as indicated on figure) by two-sided pre-ranked gene set enrichment analysis with Benjamini-Hochberg FDR correction.