Methyl-SNP-seq reveals dual readouts of methylome and variome at molecule resolution while enabling target enrichment

Abstract

Covalent modifications of genomic DNA are crucial for most organisms to survive. Amplicon-based high throughput sequencing technologies erase all DNA modifications to retain only sequence information for the four canonical nucleobases, necessitating specialized technologies for ascertaining epigenetic information. To also capture base modification information, we developed Methyl-SNP-seq, a technology that takes advantage of the complementarity of the double helix to extract the methylation and original sequence information from a single DNA molecule. More specifically, Methyl-SNP-seq uses bisulfite conversion of one of the strands to identify cytosine methylation while retaining the original four-bases sequence information on the other strand. As both strands are locked together to link the dual readouts on a single paired-end read, Methyl-SNP-seq allows detecting methylation status of any DNA even without a reference genome. Because one of the strands retains the original four nucleotide composition, Methyl-SNP-seq can also be used in conjunction with standard sequence-specific probes for targeted enrichment and amplification. We demonstrate the usefulness of this technology in a broad spectrum of applications ranging from allele-specific methylation analysis in humans to identification of methyltransferase specificity in complex bacterial communities.

Simultaneous assessment of human genome and methylome data in a single experiment using limited deamination of methylated cytosine

“…We processed these data sets and extracted methylation information using the Bismark pipeline: (1) for a fair comparison, we shortened the paired-end reads to 100 bp long and trimmed the Illumina adapters as well as the first two bases of Read2 (Trimgalore --clip_R2 2); (2) we aligned trimmed reads to the human GRCh38 genome using Bismark (version 0.22.3); (3) we filtered the PCR duplicates and incomplete bisulfite conversion using Bismark deduplicate_bismark and filter_non_conversion, respectively; (4) we combined the replicates and extracted CpG sites methylation using bismark_methylation_extractor. The methylation level of targeted regions such as CGI or promoters was calculated as explained previously (Yan et al 2022).…”

“…For NA12878 BS-tagging sequencing (mentioned as Tn5 BS in Figure) (Suzuki et al 2018), we converted its processed methylation information on hg19 to hg38 using the UCSC liftOver tool (Hinrichs et al Cold Spring Harbor Laboratory Press on June 27, 2024 -Published by genome.cshlp.org Downloaded from 2006). For NA12878 Nanopore sequencing, the data processing and methylation quantification were described in (Yan et al 2022).…”

“…To quantify how correlated the methylation levels are between RIMS-seq2 and other technologies for methylation analysis such as bisulfite sequencing or EM-seq as well as technologies that provides both sequence and methylation readout such as Nanopore, 5-letter-seq (Füllgrabe et al 2023) and Methyl-SNP-seq (Yan et al 2022), we proceed with a genome-wide comparison of various publicly available methylation datasets done on the same cell lines. Since RIMS-seq2 cannot reliably identify methylation at base resolution with current standard sequencing depth, we aim at obtaining regional aggregated methylation levels (RAML) values at defined genomic regions.…”

Simultaneous assessment of human genome and methylome data in a single experiment using limited deamination of methylated cytosine

“…(2) we aligned trimmed reads to the human GRCh38 genome using Bismark (version 0.22.3); (3) we filtered the PCR duplicates and incomplete bisulfite conversion using Bismark deduplicate_bismark and filter_non_conversion, respectively; (4) we combined the replicates and extracted CpG sites methylation using bismark_methylation_extractor. The methylation level of targeted regions such as CGI or promoters was calculated as explained previously (Yan et al 2022). For NA12878 BS-tagging sequencing (mentioned as Tn5 BS in Figure) (Yan et al 2022;Suzuki et al 2018), we converted its processed methylation information on hg19 to hg38 using the UCSC liftOver tool (Hinrichs et al 2006).…”

“…Recently, we and others have used the redundancy of the double stranded DNA to identify converted cytosine and thus, obtain DNA methylation information and genomic variants simultaneously (Yan et al 2022) (Liang et al 2021) (Füllgrabe et al 2023). While this setup allows for dual readouts in a single dataset, the experimental application of these techniques proved to be a significant departure from standard library preparation as the procedure involves linking double stranded DNA together.…”