how many replicates do you have? how large are your regions? 

if you have enough replicates then you can compare the distribution per each site across different conditions. otherwise you can group (depending from the distance) CpG sites in "island" and take the average methylation status of that region and then compare the different conditions. 

Hi Alice, for Small size, it is better to use Fisher test.

Dear Alice,

you can use of chi square or Fisher Test for your analysis,

best 

Building on Mostafa and Anter's answers:

you can build a contingency table as follow (if the layout works in this textbox):

                |-----------------------|-----------------------------|------------------------|

                 |    # methylated  |  # Non methylated    |      total                |

--------------|-----------------------|----------------------------|------------------------|

Treatm     |        34               |                23               |        57                |

--------------|----------------------|-----------------------------|------------------------|

ctr            |         66               |                55              |           121            |       

-------------|-----------------------|-----------------------------|------------------------|

total         |        100              |               78               |            178           |

-------------|-----------------------|-----------------------------|------------------------|

i made up the numbers but those are the number of 1s and 0s

then you use a chi Fisher's exact test, you can do it online here:

http://graphpad.com/quickcalcs/contingency1.cfm

or in R with the fisher.test() function.

fisher test will give you (by definition) an exact p-value, chi square gives more an approximation.

Many thanks to all! Very helpful! Will give it a try tonight!

Dear Alice

I do agree that the fisher test could be useful, but that only works if you have two groups with two possible outcomes. If you have more than two groups, or if you have more that two possible outcomes, then the chi square is the best option. If your groups are too small, you may not reach the criterea to use the chi square test. Either you can try to increase the number of your samples, or try to work with continous values. I know your data is binary, but you don't explain if that is the outcome of your test, or if you put  a threshold to make your data binary. If the case is the second, you may use your data before the transformation and try with a Mann-Withney U test, or some other, for small samples.

Dear Ernesto,

thanks for your reply. My data is binary because the program I use to analyze my sequences produces the results in that format.. I have an N of 8 per group (5 groups in total), so it seems Chi square is the way to go, unless you have any further suggestions!

Thanks for all your help

Alice

Why not try Correlation and regression analysis. Since you are looking for the relationship between two variables and your data are quantitative.

Dear Alice

Go ahead with the chi square. My only concern is if your expected frequencies meet the criteria to use the chi square test.