Skip to main content
Fig. 4 | Genome Biology

Fig. 4

From: MethylationToActivity: a deep-learning framework that reveals promoter activity landscapes from DNA methylomes in individual tumors

Fig. 4

M2A recapitulates subtype differences in RMS. a A t-distributed Stochastic Neighbor Embedding (tSNE) analysis of observed (left), M2A inferred (center: pre-transfer, right: post-transfer) H3K27ac promoter levels. Embryonal and alveolar RMS subtypes are well separated in each analysis, demonstrating that M2A inferred H3K27ac levels maintain the delineation of RMS subtypes, consistent with the observed H3K27ac tSNE analysis. b–d The subtype-specific genes GAS2 (b) and NOS1 (c) show subtype distinct patterns of DNAm, H3K27ac, and H3K4me3 levels. The windowed average DNAm feature (2500-bp windows, over the genomic region TSS ± 25 kb) is shown as example (partial) M2A input. These subtype differences were faithfully recapitulated by the M2A H3K27ac inferences for GAS2 (d) and NOS1 (e)

Back to article page