26th Oct, 2016

ecSeq Bioinformatics GmbH

Question

Asked 23rd Oct, 2014

I have followed all the standard protocols for bisulfite sequencing like cloning, etc. I got the sequence and I can identify whether the CpG is methylated or unmethylated but I don't know how to calculate the percentage of methylation from the sequence electrophoregram. How do I do this?

**Get help with your research**

Join ResearchGate to ask questions, get input, and advance your work.

22. - 25. November 2016

Leipzig, Germany

- 1.92 MBecseq_2016-7_Methylation.pdf

1 Recommendation

Yes, you are right! You have to sequence many of the clones to calculate the level of methylation (At the very least 10 clones, better 20 or more).

Then you look at the same CpG position in all seqeunces. Methylated cytosines will stay "C". Unmethylated cytosines will have been converted to "T" by the bisufilte treatment. Count the number of "C"s and "T"s at one position, and you can calculate the level of methylation.

examples:.

If 10/10 nucleotides at this position are "C", then the DNA is fully (100%) methylated at this position.

If 3/10 nucleotides at this position are "C", then the DNA was 30% methylated at this position. This is also called "partial methylation".

If 0/10 nucleotides at this position are "C", then the DNA is unmethylated (0%) at this position.

-----

To explain "partial methylation" (also called "differential methylation") a bit more:

One CpG on *one chromosome* can only be methylated or unmethylated, but not a mix. However, you have two copies of each chromosome in one cell, so *one cell *can be 0%, 50% or 100% methylated at one CpG.

When you isolate DNA, you have a DNA mix of many thousand cells. Depending on how many of the cells are unmethylated (0%), hemimethylated (50%) or fully methylated (100%) , you can get a level of methylation from 0 - 100%.

Because bisulfite sequencing has a relatively high error rate, samples between 0% and 10% are usually classified as "unmethylated", and samples between 90% and 100% are classified as "fully methylated". [But these thresholds are not fixed. Some groups use 0-5% and 95-100%, others use 0-20% and 80-100%.]

"partial methylation" refers to a state where the DNA is not fully methylated and not unmethylated, so in most cases partial methylation is between 10% and 90%. [or for the examples from other groups 5%- 95% and 20%-80%]

27 Recommendations

I use QUMA (QUantification tool for Methylation Analysis) (http://quma.cdb.riken.jp/).You can easily align, visualize and quantify

bisulfite sequence data for CpG methylation analysis. It also performes statistical analyses!

6 Recommendations

Thanks Sangalli. QUMA shows methylation levels compared to the nearby CpGs. Is it possible to calculate the methylation level for a single CpG from bisulfite sequence? If not, how much this method QUMA is reliable?

Could you be more specific on what you did, please? What did you clone and when? Did you use NGS or Sanger?

For now, it sounds to me like you did Sanger sequencing of bisulfite-treated DNA from many clones. The Sanger sequence for each clone should be a definite result: either methylated or unmethylated. There should not be a mixture of methylated and unmethylated signal in the electropherogram of one clone. If you can see a mixed signal, something most likely went wrong in the cloning procedure. In short: you should not calculate the methylation from the electropherogram.

Once you have counted the number of methylated and unmethylated clones, you then calculate the percentage of methylation. (e.g. 15 methylated clones and 5 unmethylated clones = 75 % methylation)

6 Recommendations

Thanks Wagner. I had done bisulfite specific PCR, gel eluted the specific product and cloned it. I got more than one clone but I sanger sequenced only one clone. Do I have to sequence all the clones to find out level of methylation?

I am bit confused. A CpG can be a methylated one or unmethylated one. How can it be a partially methylated. If I am looking into a single CpG, is it possible that I get partial methylation? Kindly clarify the term PARTIAL METHYLATION

Yes, you are right! You have to sequence many of the clones to calculate the level of methylation (At the very least 10 clones, better 20 or more).

Then you look at the same CpG position in all seqeunces. Methylated cytosines will stay "C". Unmethylated cytosines will have been converted to "T" by the bisufilte treatment. Count the number of "C"s and "T"s at one position, and you can calculate the level of methylation.

examples:.

If 10/10 nucleotides at this position are "C", then the DNA is fully (100%) methylated at this position.

If 3/10 nucleotides at this position are "C", then the DNA was 30% methylated at this position. This is also called "partial methylation".

If 0/10 nucleotides at this position are "C", then the DNA is unmethylated (0%) at this position.

-----

To explain "partial methylation" (also called "differential methylation") a bit more:

One CpG on *one chromosome* can only be methylated or unmethylated, but not a mix. However, you have two copies of each chromosome in one cell, so *one cell *can be 0%, 50% or 100% methylated at one CpG.

When you isolate DNA, you have a DNA mix of many thousand cells. Depending on how many of the cells are unmethylated (0%), hemimethylated (50%) or fully methylated (100%) , you can get a level of methylation from 0 - 100%.

Because bisulfite sequencing has a relatively high error rate, samples between 0% and 10% are usually classified as "unmethylated", and samples between 90% and 100% are classified as "fully methylated". [But these thresholds are not fixed. Some groups use 0-5% and 95-100%, others use 0-20% and 80-100%.]

"partial methylation" refers to a state where the DNA is not fully methylated and not unmethylated, so in most cases partial methylation is between 10% and 90%. [or for the examples from other groups 5%- 95% and 20%-80%]

27 Recommendations

You can use getMethylationStats() function from https://code.google.com/p/methylkit/

Very easy to use.

How to calculate the log2 fold change?

Question

24 answers

- Asked 7th Nov, 2017

- Ganesh Ambigapathy

I have 3 groups. 1. Control 2. Disease 3. Treatment. I want to lookup the gene expression btw these groups, compared with control (whether is upregulated or downregulated).

I did real-time qPCR and have ct values. I calculated ∆Ct = Ct[Target]-Ct[Housekeeping] ... and ∆∆Ct = (∆Exp.)-(∆Control) and got the -∆∆Ct log-fold-change. It looks all the values are almost same and not much different between the groups.

My questions are,

1. What I did was right?

2.If I plot a graph what should I mention in y-axis?

3. Is there any other better way to calculate the gene expression results better?

4.How to calculate log2 fold change and does it helps to see the results more clearer?

p.s I have attached the .xls file for your reference.

Thanks in advance...!!!

Data

- Aug 2013

Correlation between microarray log2-ratios values and bisulfite sequencing across treatment groups.
Bisulfite sequencing was performed for all sample series for MOCK, AZA, and DAC treated samples for genomic regions showing either increases or decreases in methylation after treatment. The percentage of methylation determined by bisulfite sequencing...

Data

- Jul 2016

Validation of methylation percentages for nine differentially methylated genes between Nasonia vitripennis (Nv) and Nasonia giraulti (Ng) using the PyroMark assay in independent biological replicates.
Plotted in each bar plot are CpG methylation percentages estimated from whole-genome bisulfite sequencing data (left six bars) and single gene PyroMa...

Data

- Mar 2012

Direct bisulfite sequencing of significant differentially methylated genes/genomic regions in MeDIP-seq samples. (a, b, c, d, e) For all figures, the horizontal line represents the position of each CpG investigated and the vertical line is the percentage of the methylation at particular CpG site from 0–100%. The analysis was performed using QUMA.
(...

Get high-quality answers from experts.