How to calculate conversion rate of a bisulfite sequencing experiment from lambdaDNA?

[prem]

Hi all
I am new to bisulfite sequencing analysis. Could anyone tell me how to calculate the conversion rate of a bisulfite experiment (lambda DNA has added in the genomic DNA).
It is a WGBS data. I have the sequence data.

Is it by aligning to the reference lambda genome? Can anyone help me in this?

Thanks in advance..

[pmiguel]

Hi all
I am new to bisulfite sequencing analysis. Could anyone tell me how to calculate the conversion rate of a bisulfite experiment (lambda DNA has added in the genomic DNA).
It is a WGBS data. I have the sequence data.

Is it by aligning to the reference lambda genome? Can anyone help me in this?

Thanks in advance..

Is it normal to use phage DNA to calibrate bisulfite sequencing? Seems hazardous to me. E. coli will methylate some cytosine residues. Packaging will interfere with this process. So you generally see a variable amount of cytosine methylation in phage DNA.
You would want to use phage lambda DNA from an E. coli host that had all its C-methylation pathways knocked out.
--
Phillip

[prem]

Hi Philip

I am new to this and not aware of the situation you have mentioned. Can you please explain?

[pmiguel]

Hi Philip

I am new to this and not aware of the situation you have mentioned. Can you please explain?

E. coli normally methylates its DNA. Look up "dcm methylation" for more information. phage lambda causes E. coli to produce numerous copies of itself during the lytic phase -- this is the basis of generating phage lambda DNA. Presumably because this process is rapid and the DNA is being packaged into new lambda particles, not all the sites susceptible to methylation will actually be methylated.

Of course there are E. coli strains that have all the DNA methylases knocked out. But, at least in the past, they grew poorly and so would not tend to be used for phage lambda preps unless knocking out the methylation was critical.

--
Phillip

[prem]

Thank you Phillip for the information. Anyway I would like to read more about it. I will read about "dcm methylation" also. cl857 Sam7 Lambda genome was used for alignment purpose.

[pmiguel]

Thank you Phillip for the information. Anyway I would like to read more about it. I will read about "dcm methylation" also. cl857 Sam7 Lambda genome was used for alignment purpose.

Ah, you should be okay then. That is a dcm- strain.

Well, I'm by no means an expert on doing the analysis you mention. But the idea is that 100% of the C's would ideally have been converted to T's. So you would do an alignment of your reads to phage genome sequence and check to see what percentage conversion you are seeing.

I still think using phage lambda was not a good choice for this control. If you had another DNA source that was not packaged, that was isolated from a dcm+ host, then all the dcm sites would be methylated. That way you would be able to check for 100% conversion of the non-methylated C's and 0% conversion of the methylated C's. But maybe the latter is just not an issue with bisulfite conversion?

--
Phillip

[pmiguel]

Thank you Phillip for the information. Anyway I would like to read more about it. I will read about "dcm methylation" also. cl857 Sam7 Lambda genome was used for alignment purpose.

Wait. Just to be clear -- where did you obtain this lambda DNA from? Promega, for example, provides this type of phage DNA grown in a dcm- host, here. But I see no reason this phage would not grow in a dcm+, strain.

--
Phillip

[prem]

Thank you Phillip for your time.
I am using lambda DNA provided by NEB.

Now, after reading and searching different articles, I reached a conclusion to calculate the rate. That is same as you have suggested in your previous mail.

* Align the reads to lambda genome using bismark
* Using bismark_methylation_extractor, get methylated/unmethylated count from the alignment output
* rate = [unmethylated/(methylated + unmethylated)] * 100

Please correct me if I am wrong...

[pmiguel]

Thank you Phillip for your time.
I am using lambda DNA provided by NEB.

Then you will have an issue to deal with. From the NEB site:

FAQ: Is the Lambda DNA (N3011) methylated?
Yes, it is but we do not know exactly which A and C sites get methylated since the methylation in Lambda is not 100% done. It is dam+, dcm+, EcoK- (strain C190). We do sell N6-methyladenine-free Lambda DNA (N3013), which is dam-, dcm+, EcoK+ (strain GM33).

The methylase encoded by the dam gene (Dam methylase) transfers a methyl group to the N6 position of the adenine residues in the sequence GATC (1,2). The Dcm methylase, encoded by the dcm gene, methylates the internal cytosine residues in the sequences CCAGG and CCTGG (1,3) at the C5 position. The EcoKI methylase, M.EcoKI, modifies adenine residues in the sequences AAC(N6)GTGC and GCAC(N6)GTT.

Not all DNA isolated from E.coli is methylated to the same extent. While pBR322 DNA is fully modified, only about 50% of Lambda DNA Dam sites are methylated, presumably because the methylase does not have the opportunity to methylate the DNA fully before it is packaged into the phage head.

All NEB lambda DNA is produced from a dcm+ host strain. As a result, the second C of sites CCWGG in the sequence will be resistant to bisulfite conversion as some of them will be methylated.

This should be easily dealt with. You could ignore the conversion rate of C's at these sites and only consider other C's in your calculation.

--
Phillip