Protocol Online logo
Top : New Forum Archives (2009-): : DNA Methylation and Epigenetics

statistical analysis or quantitative analysis method for bisulfite sequencing r - (Oct/01/2011 )

Hollow everyone~

I am new here and have some question need help

In my study, I want to investigate the CpG island methylation pattern of interesting gene between tumor and normal cells.

So,I had collected 60 paired clinical DNA (tumor part and normal part) and performed Bisulfite PCR -->TA clone --> picked 10~12 clone for Sequencing

Now, I got all the sequencing result and aligned them by BiQ analyzer

The length of my interesting region is about 600bp, 22~26 CpG sites were located in this region.

The questions are following:

1. What quantitative methods should be used for calculated the degree of methylation level?

example 1:
if "A" pateint, tumor part contain 22 CpG sites and 10 clones were be calculated -> Totally, 22*10= 220 CpG sites act as denominator, between them,165 CpG sites(75% of 220 CpG sites) display unmethylated pattern. Than, We called this patient' tumor part in the CpG sites display hypomethylation pattern.

example 2:
if "B" patient, tumor part also contain 22 CpG sites and 10 clones were be calculated-->
7 of 10 clone display 75% unmethylated pattern (17 CpG site/22 CpG sites). han, We called this patient' tumor part in the CpG sites display hypomethylation pattern.

I just confused for which method is more suitable for analyze the bisulfite sequencing data in my study? and how to decide the parameter or cut off value? (60%, 70%, or even 90%)

Does there are any software sutable for this work?

2. Actually, My boss asked me to used another method to verify my bisulfite-sequencing results. the method had better can more easily quantify the results between normal and tumor between different patient .

Does there are anyone can give me an comment for my following experiment?

I will deeply appreciate your help




Wusian, BiQ Analyser can give you these metrics. In reality, there is no one way to represent your data, how to average etc etc.

As for an alternative method, I suspect your PI is suggesting something like pyrosequencing or SEQUENOM.


2. The pyrosequencer will give you a quantification value but it's based on how efficient your PCR reaction is.


How can quantify the methylation status just sequencing the PCR product without clonning? Should I performe MSP? Or should I consider the result just as methylated and non-methylated? Complicated? Pyrosequencing would work better?


Lilip, depends on what you want to look at. the methods you state actually measures slightly different things.