HomeBiotechnologyExcessive-throughput sturdy single-cell DNA methylation profiling with sciMETv2

Excessive-throughput sturdy single-cell DNA methylation profiling with sciMETv2

Moral assertion

Human tissue used on this research was obtained as de-identified specimens from the Oregon Mind Financial institution which is overseen by the OHSU Institutional Overview Board and consented for genetic information sharing and genomic information beneath restricted entry for analysis functions solely. Uncooked information from specimens are deposited within the Neuroscience Multi-omic Information Archive (NeMO) beneath restricted entry with processed information that doesn’t comprise genomic variants overtly accessible. Mouse specimens have been obtained from discarded tissue and didn’t require moral approval.

Statistics and reproducibility

No statistical technique was used to predetermine pattern measurement. Cell deposition for second-tier indexing was decided primarily based on empirical information and that produced a <5% collision price for the sciMETv1 assay. All cells assembly a minimal protection threshold and beneath 10% methylated CH bases have been included on this research. The experiments weren’t randomized. The Investigators weren’t blinded to allocation throughout experiments and consequence evaluation.

sciMETv2 nucleosome disruption and tagmentation

An in depth step-by-step protocol is supplies as Supplementary Word 1 within the Supplementary Info related to this manuscript.

Human center frontal gyrus samples have been obtained from the Oregon Mind Financial institution the place they have been saved by instantly inserting dissected tissue cassettes in a −80 °C freezer. Mouse (C57BL6, feminine, P56) specimens have been obtained as discarded tissue and preserved by flash-freezing in a liquid nitrogen bathtub. A portion of tissue was remoted utilizing a sterile razor blade and positioned in a dish containing ice chilly Nuclei Isolation Buffer (NIB, as in Thornton et. al. 2021; obtainable as ScaleBio Half. No. 230041)17. The slurry was then homogenized with a Dounce homogenizer utilizing a free pestle for 7 strokes, incubating 10 min on ice, and repeating the homogenization with a decent pestle for 7 strokes. Nuclei have been spun down by centrifugation at 500 xG for five min at 4 °C, the supernatant eliminated, after which resuspended in 200 µl ice chilly NIB. Nuclei have been then carried by means of nucleosome disruption in response to Mulqueen et. al. 2021 utilizing buffers obtainable from ScaleBio12. Nuclei have been first fastened at room temperature in 1 million nuclei aliquots utilizing 1 mL Fixation Response Mastermix (FRM, ScaleBio Half No. 230011) for every aliquot. Nuclei have been incubated at room temperature for 10 min with occasional mixing. Fixation was stopped by including an equal quantity of Cease Grasp Combine (STP, ScaleBio Half No. 230021) and incubated on ice for five min. Nuclei have been spun down at 500xG at 4 °C for five min and resuspended in Nucleosome Depletion Mastermix (NDM, ScaleBio Half No. 230031) and incubated at 37 °C for 20 min. Reactions have been then spun down at 500xG at 4 °C for five min and resuspended in 200 µL NIB. Aliquots have been mixed at this level.

Tagmentation was carried out in a 96-well plate. Every properly contained 5000 nuclei, 2.5 uL 4X-Faucets-TD (obtainable from ScaleBio previous to business availability; 33 mM TAPS pH = 8.5 [Sigma, Cat. T5130], 66 mM KOAc [Sigma, Cat. P1190], 10 mM MgOAc [Sigma, Cat. M5661], 16% DMF [Sigma, Cat. D4551]), 2 uL 5 µM Tn5 complexes with methylated Cs utilizing the oligo sequences detailed in Supplementary Information 1 (obtainable from ScaleBio upon request previous to business availability) and sufficient NIB to fill the quantity to 10 uL. Word the 4X-Faucets-TD is to be made recent, the day of the experiment. The plate was then incubated for 15 min at 55 °C. Afterwards it was placed on ice. For the preparation leveraging 288 tagmentation indexes, three 96-well plates have been ready identically however with using three completely different units of 96 barcoded Tn5 complexes.

All wells have been pooled and so they have been run by means of a 40 µm cell strainer. 3 uL 5 mg/mL DAPI was added to stain the nuclei. They have been FACS/FANS sorted at 15–30 nuclei per properly into 96-well plates with every properly containing 1 uL M-Digestion Buffer (Zymo Analysis: Cat. D5021-9), 0.07 µL reconstituted Qiagen Proteinase Ok (Qiagen: Cat. 19131), and 0.93 µL dH2O. These post-sort plates have been then spun down at 500xG at 4 °C for 10 s. They have been then incubated at 50 °C for 20 min.

sciMETv2 bisulfite conversion

Bisulfite conversion was tailored from the snmC-seq2 protocol6, 1 bottle of CT Conversion Reagent was reconstituted with 7.9 mL M-Solubilization Buffer and three mL M-Dilution Buffer. The bottle was shaken vigorously to dissolve. After the reagent was dissolved, 1.6 mL M-Response Buffer was added. After a fast spin down of the post-sort plates, 15 µL of the reconstituted CT Conversion Reagent was added to every properly. The plates have been then incubated at 98 °C for 8 min, then at 64 °C for 3.5 h with a closing 4 °C maintain in a single day.

To scrub the bisulfite reactions 80 µL M-Binding Buffer was added to every properly and your entire quantity of every properly was then transferred to a 96-well Zymo-Spin I-96 Plate (Shallow properly). The column plates have been then spun down at 2200 xG for 8 min. Circulate-through was discarded and 100 µL M-Wash Buffer was added to every properly and spun at 2200 xG for 8 min or till all Wash Buffer got here out. Circulate-through was discarded and 50 µL M-Desulphonation Buffer was added to every properly. The plates have been incubated for 15 min at room temperature. Afterwards, they have been spun at 2200 xG for 8 min, discarding the flow-through. 200 µL M-Wash Buffer was added to every properly and spun out on the above velocity and time. After discarding the flow-through, the plates have been spun once more to fully dry the columns. Bisulfite transformed DNA was eluted by including Buffer EB that was preheated at 55 °C.


Bisulfite transformed DNA was eluted by including 25 µL Buffer EB to every column. The plate was incubated for 4 min at 55 °C by inserting on high of a 96-well plate containing the next in every properly: 17.8 µL dH2O, 5 µL 10X NEB Buffer 2.1, 2 µL 10 mM dNTPs and 0.2 µL 100 µM 9H Random Primer. The assembled plates have been spun on the above velocity and time to elute the DNA. That is the linear amp (LA) plate.

The LA plate was then warmth shocked at 95 °C for 45 s after which positioned on ice till cool. 10 items of Klenow exo was added and positioned on a thermocycler with the next incubations: 4 °C for five min, ramp of 1 °C each 15 s till 37 °C is reached and incubate for 90 min at 37 °C, then maintain at 4 °C. That is the primary LA cycle.

The plate was then warmth shocked once more at 95 °C for 45 s and positioned on ice. The next was then added to every properly of the plate: 0.1 µL 100 µM 9H Random Primer, 1 µL 10 mM dNTPs, 0.5 µL 10X NEB Buffer 2.1, 1.65 µL dH2O and a couple of µL of Klenow exo- (5U/uL). The above thermocycling protocol was run once more. That is the second LA cycle. This cycle was repeated twice extra for a complete of 4 LA cycles.

After a 1.1:1 (SPRI to template ratio) SPRI cleanup, the plate was eluted with 21 µL Buffer EB right into a plate containing the next (closing quantity is 50 µL): 25 µL Q5U 2X Grasp Combine, 2 µL 10 µM TruSeq i5 primers, 2 µL TruSeq i7 primers and 0.5 µL EvaGreen 100X. The plate was positioned on a qPCR machine utilizing the next program: 95 °C 2 min preliminary activation, 94 °C 80 s denaturation, 65 °C 30 s annealing, 72 °C 30 s extension, 72 C 10 s plate learn. As soon as most wells started to plateau, the plate was pulled. All wells have been pooled for cleanup and library QC.


Bisulfite transformed DNA was eluted by including 5 µL preheated Buffer EB to every column within the plate. The assembled plates have been incubated for 4 min at 55 °C by inserting on high of a 96-well plate containing 1 µL 20 ng/uL ET-SSB.

The plate was then warmth shocked at 95 °C for 3 min and positioned on ice for two min. 1 µL 0.75 µM pre-annealed P5 adapter was then added to every properly. The ligation grasp combine (4 µL 50% PEG 8000, 0.75 µL SCR Buffer, 0.1 µL 1 M DTT, 0.1 µL 100 mM ATP, 0.125 µL T4 PNK [10,000 U/mL, NEB], and 0.125 µL T4 Ligase [2,000,000 U/mL, NEB]) was added to the plate at room temperature. Word: The splint ligation response depends on the ligation grasp combine being arrange at room temperature and was carried out in response to Kapp et al. 2021 protocols13. No cleanup was carried out between ligation and PCR.

For indexing PCR, the next was added to every properly of the ligation plate: 10 µL 5X VeraSeq GC Buffer, 2 µL 10 µM dNTPs, 1.5 µL VeraSeq Extremely Enzyme, 24 µL dH2O, 0.5 µL 100X EvaGreen, 1 µL 10 µM TruSeq i5 primers, 1 µL 10 µM TruSeq i7 primers. The next biking program was used: 98 °C 30 s preliminary activation, 98 °C 10 s denaturation, 57 °C 20 s annealing, 72 °C 30 s extension, 72 °C 10 s plate learn. As soon as most wells started to plateau, the plate was pulled. All wells have been pooled for cleanup and library QC.

Computational evaluation

All code used to course of and analyze information could be discovered right here: github.com/adeylab/sciMETv2, apart from the preliminary demultiplexing which could be discovered right here: github.com/adeylab/unidex, or different wrapper capabilities right here: github.com/adeylab/scitools.

Learn processing and alignment

Uncooked sequence reads have been demultiplexed utilizing unidex (github.com/adeylab/unidex) with a customized mode that makes use of the three indexes: ahead PCR, reverse PCR, and tagmentation index, permitting for a hamming distance of two. The primary 10 bases of learn 1 (random priming area), bases 1-8 (index) and 9-29 of learn 2 (mosaic finish) have been all trimmed throughout this processing. Reads have been then carried by means of sciMET_trim.pl, part of the sciMET GitHub repository (github.com/adeylab/sciMETv2), utilizing ‘-e 10’ which makes use of trimmomatic18 to trim adapter sequences in single-end mode, adopted by eradicating the final 10 bases of learn 2 since they might embrace the random primer area. Reads have been then aligned utilizing bismark (v0.23.1)19 with the wrapper script: sciMET_align.pl which makes use of the ‘–pbat’ choice for learn 1 and the default directional choice for learn 2. Preliminary alignment was to a reference genome containing each the human and mouse genome to determine species identification and assess doublet charges and the presence of any cross-cell contamination utilizing scitools barnyard-compare (github.com/adeylab/scitools)20 on the bam file after barcode-aware duplicate learn elimination utilizing sciMET_rmdup.pl. As soon as species have been recognized, trimmed reads for these cells have been then break up to the respective species-specific learn information utilizing sciMET_speciesSplit.pl after which aligned to their respective genomes following the identical workflow.

For the preparation that leveraged 288 tagmentation indexes (3 × 96-well plates) we optimized the processing and alignment pipeline with up to date scripts, additionally obtainable within the sciMET GitHub repository. The brand new trimming workflow makes use of TrimGalore (v0.6.5), which leverages CutAdapt (v4.1)21, inside the wrapper script sciMET_trim.pl, which performs a pair-wise learn trim. Correctly paired, trimmed reads have been then aligned utilizing bismark19 with the wrapper script sciMET_align_pe.pl which swaps reads 1 and a couple of after which performs a paired alignment utilizing the ‘–native’ choice. We discovered that these variants supplied a notable enchancment in processing time and reaching a comparable alignment price.

TSS Enrichment was calculated utilizing averaged 200 bp home windows 1000 bp away from TSSs as background and the 200 bp centered on the TSS as sign. Shotgun WGS information produces a worth close to 1 (0.944) utilizing tagmentation-based shotgun sequencing of the human GM12878 cell line versus different TSS Enrichment strategies that usually produce values a lot greater. This technique was chosen as tagmentation of intact nuclei with out nucleosome disruption (i.e. ATAC-seq) produces excessive TSS Enrichment and the purpose of nucleosome disruption is to ablate this sign and obtain a worth at or beneath 1.

Methylation calling and evaluation

Species-specific aligned and duplicate eliminated bam information have been processed utilizing the wrapper: sciMET_extract.pl which makes use of the bismark methylation extraction instrument with the next choices: ‘–complete–merge_non_CpG–single-end–no_header’. The extracted information have been then break up into particular person chromosome information to assist in processing time and reminiscence for future operations with a separate folder for CG and CH methylation calls. The CH methylation calls have been then used to assemble a mCH matrix of cells × 250 kbp genomic window utilizing sciMET_meth2mtx.pl. The ensuing matrix was filtered to first solely embrace cells with a minimal of 75% of home windows with 20 or extra known as CH websites adopted by filtering home windows to solely retain these with not less than 75% of passing cells with 20 or extra known as CH websites utilizing sciMET_filtMtx.pl. The matrix was then taken by means of SVD, computing 50 elements utilizing the irlba operate in R inside the scitools irlba operate. This matrix was then used for Louvain-based clustering with scitools matrix-pg, a wrapper for the phenograph instrument22, in addition to used for visualization utilizing UMAP23.

Protection per-cell for every technique was assessed by performing 100 iterations of random cell choice the place for 250 cells the cumulative methylome protection was assessed. This was carried out for every technique and a downsampled model of the linear amplification technique assessed at a comparable variety of uncooked sequence reads to the splint ligation libraries.

Promoter methylation for genes by clusters was decided utilizing the sciMET script: getGeneMeth.pl with the cluster identities beforehand decided for the complete set of genes in RefGene taking the window 5 kbp upstream and a couple of.5 kbp downstream of the TSS. Gene physique CH methylation ranges have been calculated utilizing the identical script however assessing your entire gene physique together with 2 kbp upstream and downstream. These values have been aggregated over annotated clusters after which z-scored throughout clusters to provide plots in Fig. 2c.

Cell-cell similarity scores have been calculated primarily based on the technique detailed in Hui et. al. (2018)5 the place overlapping CG websites between cells restricted to regulatory areas (promoters encompassing +1.5 and −1 kbp from TSS) have been assessed for shared methylation standing utilizing the sciMET_pairwise.pl script. Imply cross-cluster cell similarities have been hierarchically clustered utilizing R heatmap.2.

Motif methylation was assessed by figuring out all genomic loci inside the GRCh38 reference genome utilizing Homer (v3.0)24. Home windows have been then generated in 10 bp increments out to 500 bp from the middle of the motif utilizing the sciMET_featuresToScanBed.pl script, and the imply methylation worth calculated for every cluster utilizing both CG or CH methylation calls utilizing the sciMET_getWindowMeth.pl script.

snmC-seq2 information was downloaded from the NCBI Gene Expression Omnibus utilizing accession “GSE112471”. Reads have been processed to take away adapters and brief fragments utilizing cutadapt21 as detailed within the snmC-seq2 workflow6 after which aligned utilizing our processing workflow detailed above, swapping reads 1 and a couple of to account for the reverse adapter configuration of snmC-seq2 vs sciMETv2. PCR duplicates have been eliminated and methylation calls carried out utilizing the identical workflow as detailed above. The ensuing CH methylation matrix was then merged with the sciMETv2 built-in matrix previous to filtering and processing.

CG-centered analyses have been carried out as within the CH evaluation however utilizing 50 kbp non-overlapping home windows and filtering the ensuing matrix to retain cells with not less than 25% of home windows with 5 or extra CG websites lined and home windows with not less than 50% of cells passing the identical protection threshold.

Reporting abstract

Additional info on analysis design is accessible within the Nature Portfolio Reporting Abstract linked to this text.



Please enter your comment!
Please enter your name here

Most Popular

Recent Comments