Extended Data Fig. 1: Library statistics for scEpi2-seq libraries and comparison to other single-cell methods.
a. Overview of data extraction strategy for scEpi2-Seq data. Every fragment resulting from MNase digestion is ligated to an adaptor containing a UMI, single-cell barcode, Illumina handle and T7 promoter. Every fragment is uniquely labeled by cell barcode and UMI, and mapping positions of R1 and R2 can be used to distinguish between duplicates generated during in vitro transcription (IVT) or PCR. In addition, aggregating all sequencing reads derived from the same DNA fragment into a consensus molecule can mitigate sequencing errors and makes methylation estimates more robust. Created in BioRender (2025) https://BioRender.com/0dai1w3. b. Mapping, Demultiplexing (that is, reads with correct cell barcode) and Deduplication (that is, removed reads with same cell barcode, UMI and position in the genome per single cell) rates per plates (indicated as dots, n = 7 independent experiments) for all experiments in this study. The boxplot is defined by the median ± interquartile range (IQR) and whiskers represent 1.5× IQR c. Mapping rates from the RPE-1 hTERT plates, which were either TAPS converted or unconverted prior during library preparation. The boxplot is defined by the median ± interquartile range (IQR) and whiskers represent 1.5× IQR. d. Comparison of mismatch rates from the RPE-1 hTERT plates, which were either TAPS converted or unconverted prior during library preparation. Only nucleotides in a non-CpG context were included. The boxplot is defined by the median ± interquartile range (IQR) and whiskers represent 1.5× IQR e. Conversion efficiency of in vitro methylated lambda phage genome (spike-ins) after TAPS conversion. Each dot represents a single 384-plate (technical replicates) and results are grouped by sample of origin (n = 7 biological replicates over samples of origins). Inset displays the same data but with a selected range of conversion efficiency (that is, 85-100%). The boxplot is defined by the median ± interquartile range (IQR) and whiskers represent 1.5× IQR f. Comparison between scEpi2-seq (this study) and other single- and multi-omic approaches for measuring histone modifications in single cells.
