Whole-Genome Bisulfite Sequencing (WGBS) Library Preparation from 100 ng Genomic DNA Using Post-Bisulfite Adaptor Tagging in HCT116 Cells

Two genotypes (parental HCT116, DNMT1/-/DNMT3B-/- DKO) each in biological triplicate (n=3 independent culture passages), 6 libraries total. Lambda DNA (unmethylated) is spiked at 0.5% w/w into every sample as a conversion control. Each library is sequenced to ~15x genome coverage (150 bp paired-end). Libraries are randomized across two flow-cell lanes to avoid lane/genotype confounding. Operators are blinded to genotype during library QC by relabeling tubes with random 4-digit codes. Target final library molarity is 4 nM for pooling.

BSL-1 for HCT116 culture. Sodium bisulfite and the concentrated CT reagent are irritants (corrosive at high pH after desulphonation) — wear nitrile gloves, lab coat, and safety glasses; handle in a fume hood when aliquoting powders. Ethanol is flammable; keep from ignition sources. Collect guanidine-containing column flow-through as hazardous chemical waste (never mix with bleach — releases toxic gas). Dispose of biological waste per institutional guidelines.

Positive control: a fully CpG-methylated genomic DNA standard (M.SssI-treated; Zymo D5014) included as a separate library — should sequence >97% methylated at CpG. Negative/conversion control: 0.5% unmethylated lambda spike in every sample — non-conversion rate must be <1% (i.e., >99% C→T at lambda cytosines). No-template control through library prep to detect adaptor-dimer/index contamination. Parental HCT116 serves as the biological reference genotype against which DKO is compared.

Bisulfite conversion efficiency >99% (lambda non-conversion <1%; CHH methylation <1% genome-wide). Parental HCT116 mean CpG methylation 70-75%; DKO <10%. Library insert size 250-600 bp with a mode near 350 bp. Mapping efficiency 60-75% (lower than standard WGS due to reduced strand complexity). Duplication rate <20% at 11 PCR cycles. Bisulfite-specific GC bias visible as a leftward GC-content shift in FastQC.

• Genomic DNA: 100 ng per sample, A260/280 1.8-2.0, high-molecular-weight (>30 kb on TapeStation), Qubit dsDNA HS quantified
• Unmethylated lambda DNA (Promega D1521), 0.5% w/w spike-in
• Bisulfite conversion kit (EZ DNA Methylation-Gold, Zymo D5005) with CT Conversion Reagent and M-Desulphonation Buffer
• PBAT-compatible random-priming oligos (oligo 1: biotin-adaptor with 9N tail; oligo 2: second-strand adaptor with 9N tail)
• Klenow exo- polymerase (NEB M0212, 5 U/µL)
• Streptavidin magnetic beads (Dynabeads MyOne C1, Thermo 65001)
• KAPA HiFi Uracil+ DNA polymerase (Roche/KAPA KK2801) for final amplification
• AMPure XP beads (Beckman A63881)
• Indexed PCR primers (8 bp unique dual indices)
• 10 mM Tris-HCl pH 8.0 (low-EDTA EB), 80% ethanol freshly prepared, nuclease-free water

To produce strand-specific, single-base-resolution DNA methylation maps across the entire genome of HCT116 cells by chemically converting unmethylated cytosines to uracil with sodium bisulfite, then sequencing on an Illumina platform. The protocol uses a post-bisulfite adaptor tagging (PBAT) strategy to tolerate the DNA fragmentation that bisulfite treatment causes, enabling robust library construction from 100 ng inputs where conventional pre-bisulfite ligation fails due to adaptor loss.

1. Combine 100 ng genomic DNA with 0.5 ng lambda DNA in 20 µL nuclease-free water in a PCR tube.
2. Add 130 µL CT Conversion Reagent; thermocycle: 98°C 8 min, 64°C 3.5 h, hold 4°C (≤20 h).
3. Bind to Zymo-Spin IC column, desulphonate 15 min RT with M-Desulphonation Buffer, wash twice, elute in 12 µL EB pre-warmed to 56°C.
4. First-strand synthesis: add 1 µL oligo 1 (10 µM), denature 95°C 1 min, snap-cool on ice; add Klenow exo- mix, ramp 4°C→37°C at 1°C/15 s, hold 37°C 90 min.
5. Capture biotinylated products on 10 µL washed C1 streptavidin beads, 20 min RT rotation; wash twice with B&W buffer; resuspend in 10 µL water.
6. Second-strand synthesis on-bead: add oligo 2 (10 µM) + Klenow exo-, same ramp/incubation as step 4; wash beads.
7. Amplify on-bead with KAPA HiFi Uracil+ and indexed primers: 95°C 2 min; 11 cycles of 95°C 20 s, 60°C 30 s, 72°C 40 s; 72°C 5 min.
8. Two-sided AMPure XP cleanup (0.6x then 0.8x) to retain 250-600 bp; elute in 20 µL EB.
9. Quantify by Qubit HS, size by TapeStation D1000, qPCR-quantify with KAPA Library Quant; normalize to 4 nM and pool.

Independent variable: genotype (parental vs DNMT1/3B DKO). Dependent variables: per-CpG methylation fraction (β = methylated reads / total reads), genome-wide mean CpG methylation, and counts of differentially methylated regions. Controlled variables: DNA input mass (100 ng), bisulfite conversion temperature/time, PCR cycle number (11), sequencing depth target (15x), and lambda spike fraction (0.5%), all held constant across samples.

We hypothesize that DNMT1/3B double-knockout HCT116 cells will exhibit a genome-wide reduction in mean CpG methylation from approximately 70-75% (parental) to below 10%, with the largest absolute losses concentrated at gene-body and intergenic CpGs while CpG-island promoters (already hypomethylated) change least. Non-CpG (CHG/CHH) methylation is expected to remain below 1% in both genotypes, serving as an internal bisulfite-conversion benchmark.

Trim adaptors and low-quality bases with Trim Galore (--paired). Align to a bisulfite-converted reference (hg38 + lambda) using Bismark with Bowtie2. Deduplicate (deduplicatebismark), then extract per-cytosine methylation calls (bismarkmethylation_extractor) and generate CX reports. Compute β-values per CpG and call differentially methylated regions with methylKit or DSS (logistic regression, minimum 10x coverage per CpG). Visualize coverage/methylation in IGV; summarize with MultiQC.

Low library yield: increase input verification (re-Qubit), confirm desulphonation buffer was added, and raise amplification to 13 cycles. High lambda non-conversion (>2%): incomplete denaturation — ensure fresh CT reagent and verify the 98°C 8-min denaturation step. Adaptor dimers (~120 bp peak): repeat 0.8x AMPure cleanup and reduce primer concentration. Low mapping efficiency (<50%): check for over-fragmentation (bisulfite over-incubation) — shorten the 64°C step to 3 h. Uneven coverage/high duplication: reduce PCR cycles and increase input DNA to 200 ng.

For each CpG with ≥10x coverage in all replicates, test parental vs DKO with DSS's Wald test on the beta-binomial model (n=3 per group). Control false discovery with Benjamini-Hochberg at q<0.05; define DMRs as ≥3 significant CpGs within 100 bp and ≥25% absolute methylation difference. Power: with n=3 and ~15x coverage, the design detects ≥20% methylation differences at >0.8 power for CpGs with ≥10x depth. Report effect sizes as Δβ with 95% CIs.