Skip to main content
. 2022 May 23;40(10):1520–1527. doi: 10.1038/s41587-022-01307-0

Extended Data Fig. 4. Learning methylation-aware binding models from EpiSELEX-seq data.

Extended Data Fig. 4

(a) Alphabet used to represent normal and methylated base pairs. (b) Same as Extended Data Figure 2a, but showing the combinations of ATF4, CEBPγ, and normal and methylated DNA that were included in each experiment and the resulting complexes that were modeled. (c) K-mer enrichment analysis for the observed ATF4 EpiSELEX-seq read counts (left), the counts predicted by a mononucleotide-only model (middle), and the counts predicted by a mono- and di-nucleotide model (right). Each scatterplot compares the 8-mer enrichment observed in the normal (x-axis) and methylated (y-axis) libraries. Every point represents an 8-mer and is colored according to the legend; color is assigned based on a 6bp matching substring between the 8mer and the IUPAC code.