A comprehensive DNA methylation atlas for noncancer human tissue types

Deciphering the tissue origin of cfDNA can reveal abnormal cell death because of diseases, which has great clinical potential in disease detection and monitoring. Here we present one of the largest comprehensive and high-resolution methylation atlas based on Reduced Representative Bisulfite Sequencing (RRBS) data of 521 noncancer tissue samples spanning 29 major types of human tissues. We systematically identified fragment-level tissue-specific methylation patterns and extensively alidated the methylation signature atlas in independent methylation datasets, orthogonal epigenomic markers, and transcription regulatory elements. Based on the rich tissue methylation atlas, we develop the first supervised tissue deconvolution approach, a deep-learning-powered model, cfSort, for sensitive and accurate tissue deconvolution in cfDNA. The RRBS libraries of the genomic DNA from the 521 tissue samples were constructed following the standard RRBS protocol. 100-200 ng of intact genomic DNA in the volume of 21.5ul was used as input material. Restriction digestion was done with 2.5ul 10xCutSmart buffer and 1ul MspI (NEB) for 18h at 37 oC and 20min at 65 oC. 0.5ul 10xCutSmart buffer, 0.3ul dACGTP mixture (100mM dATP, 10mM dCTP, 10mM dGTP), 1ul Klenow (exo-, 5U/ul, NEB) and 2.6ul RT-PCR water, 0.6ul 50mM DTT (ThermoFisher) was added to the mixture for end repair and A-overhang addition with the program 30 oC for 20min, 37 oC for 1h and 75 oC for 20min. Adapter ligation was then performed with 1ul 10xThermoFisher HC T4 ligase buffer, 0.4ul 100mM ATP (ThermoFisher), 0.2ul 50mM DTT, 1ul ThermoFisher HC T4 DNA ligase (30 Weiss Unit/ul), 30ng home-made duplex UMI adapter with all the cytosines methylated (protocol adopted from Kennedy et al.) at 16 oC for 20h and 65 oC for 20min. Bisulfite conversion of the adapter-ligated product was carried out with QIAGEN EpiTect plus DNA bisulfite kit following their protocol for two rounds of conversion. The converted product was purified with Qiagen MinElute spin column and eluted with 20ul RT-PCR water. PCR amplification was done using the NEBNext Multiplex Oligos for Illumina (2.5ul of universal and index primer each) and 25ul KAPA HiFi HotStart Uracil+ ReadyMix (Roche) with the following cycling conditions: 98 oC for 45s, 9 cycles of 98 oC for 15s, 60 oC for 30s and 72 oC for 30s, followed by a final extension at 72 oC for 5min. The PCR product was purified with 1x AmpureXP beads and eluted with 30ul EB buffer. DNA concentration was measured by Qubit 1xdsDNA HS assay. 5% TBE-UREA PAGE and bioanalyzer assay was performed as quality control on each library before sequencing. ***Submitter declares that the raw data have been deposited in EGA (EGAS00001007213) due to patient privacy concerns***