10 Ways to Improve Your Bisulfite Sequencing Results

by on 18th of January, 2013 in Next Generation Sequencing
header image copy

The importance of epigenetics in biology is increasingly acknowledged (if you’re not convinced yet, read my crash course). One commonly studied epigenetic mark is CpG methylation: cytosines that are directly followed by a guanine nucleotide (indicated by CpG), can be methylated, unlike non-CpG Cs. Since attachment of a methyl group to a cytosine can affect gene expression, you might be interested in CpG methylation at your favourite genomic locus.

If you want to know the methylation status of all individual CpG nucleotides in a genomic region of interest, sequencing of bisulfite-converted DNA is the way to go. Bisulfite is a chemical that converts all unmethylated C nucleotides in a DNA sample to Ts, while leaving methylated Cs untouched. So, if you sequence bisulfite-converted DNA, finding a C means that this residue was methylated in the non-converted DNA template (because all unmethylated Cs should have been transformed into Ts). You can then compare the converted sequence to the non-converted (original) sequence to determine which CpGs were methylated.

In my experience, depending on the locus of interest (roll up your sleeves for C/G-rich areas), bisulfite sequencing doesn’t always give perfect results the first time. The following tips might help you to get some good data quickly.

1. Design your primers well. Since you don’t know the methylation outcome of your sample beforehand, you need to make sure that primer annealing is not affected by the methylation status of your primer target sequences. Thus, you need to exclude CpGs from your primer sequences. Including non-CpG Cs is crucial because they prevent unconverted, genomic DNA contamination in your PCR product. If conversion is not 100% and you would not have selected for converted sequence through the choice of your primers, you might mistake Cs in the final sequences for methylated Cs, while they were wrongly amplified. Free software exists to help you design primers for bisulfite-converted DNA.

2. Use good quality DNA. To get clean genomic DNA from mouse tissues I used Qiagen’s DNeasy Blood&Tissue kit (ref. 69504), including an RNase treatment, and checked its quality on gel.

3. Obtain consistent conversion. I initially learnt a very long, tedious and tricky protocol, which my tutor swore by, despite all her warnings about what could easily go wrong in each step. Although this was an insightful experience, for my own experiments I’ve always used Qiagen’s Epitect  Bisulfite Kit (ref. 59104), which is a very simple protocol, with which I obtain consistent results. After conversion, the DNA is single-stranded and fragile, so repeated freeze-thaw cycles should be avoided: instead, proceed to PCR directly and aliquot the remainder of your converted DNA.

4. Use semi-nested PCR on converted DNA. PCR of bisulfite-converted DNA is less efficient than regular PCR. Hence, you will most likely need two rounds of PCR to obtain enough PCR product. Taking a semi-nested approach is a good idea. I use 4 ul of product from the first PCR reaction as template for the second (semi-nested) rePCR. Increasing the annealing temperature by 2°C for the rePCR will improve specificity.

5. Use multiple controls to monitor conversion and amplification. In addition to the obvious non-template control, I always include primers directed at the bisulfite-converted Igf2r gene, which give a clear band when conversion worked well. You could also perform PCR on a known methylated genomic region, or use an X chromosomal gene (which will be subject to imprinting, so each cell contains one methylated allele and one unmethylated allele). In this case you’d have to adapt your primer strategy to include some methylated, or unmethylated CpGs, to only amplify one allele per PCR reaction.

6. Maximize DNA yields. You might need to run ~3 parallel reactions of the rePCR to obtain sufficient material for sequencing.

7. Purify the amplification products. I obtain the highest yields by snap-freezing my cut-out bands in a column (Millipore DNA gel extraction kit, ref. LSKGel050) and spinning down the eluate, followed by classical EtOH precipitation of the DNA.

8. Subclone your bisulfite-converted PCR products. Since your starting material may contain a mixture of CpG methylation profiles, sequencing the PCR products directly will likely yield a messy electropherogram. Subcloning your PCR product gives clean results representing individual cells.

9. Choose the right purification method for sequencing. Running a standard sequencing reaction on PCR products from converted DNA may yield disappointing results. Talk to your sequencing facility about the fact that your DNA is not normal (it will hardly contain any Cs), and ask for their recommendations – you’d be surprised what a difference this makes!

10. Analyze your data carefully, with a little help. Manual inspection of the obtained electropherograms is a crucial but tedious job, so I liked to use all the help I can get. I found a great partner in the free tool BiQAnalyzer. This program aligns sequences to the target region, performs quality controls (e.g. conversion rate and homology with target sequence) and suggests when data should be excluded. It produces publishable lollipop diagrams and more. Of course, you need to keep checking what is going on, but it makes life easier.

Do you know of any other tips (or guarantees, preferably!) for successful bisulfite sequencing??

Photo Credits

Leave a Reply