Question: Meaning of the value column in bumphunter output (minfi package)
0
2.1 years ago by
Antonio M.0 wrote:

Hi everyone.

In the documentation, the value column in minfi::bumphunter output is the average difference level of methylation in the bump. My question is, does this mean that the methylation levels are greater in for example, controls? (If we are comparing control vs case) .

How can I know the meaning of this measure in this case?

Thank you all for your attention

minfi bumphunter • 814 views
modified 15 months ago by src30 • written 2.1 years ago by Antonio M.0
Answer: Meaning of the value column in bumphunter output (minfi package)
0
2.1 years ago by
United States
Kasper Daniel Hansen6.4k wrote:
That depends on your design matrix On Wed, May 17, 2017 at 10:28 AM, Antonio M. [bioc] < noreply@bioconductor.org> wrote: > Activity on a post you are following on support.bioconductor.org > > User Antonio M. <https: support.bioconductor.org="" u="" 13072=""/> wrote Question: > Meaning of the value column in bumphunter output (minfi package) > <https: support.bioconductor.org="" p="" 96046=""/>: > > Hi everyone. > > In the documentation, the value column in minfi::bumphunter output is the > average difference level of methylation in the bump. My question is, does > this mean that the methylation levels are greater in for example, controls? > (If we are comparing control vs case) . > > How can I know the meaning of this measure in this case? > > > > Thank you all for your attention > > ------------------------------ > > Post tags: minfi, bumphunter > > You may reply via email or visit Meaning of the value column in bumphunter output (minfi package) >

I have this.

g <- as.factor(c(rep("c", 17), rep("t", 40))) # c for controls vs t for tumors
matriz <- model.matrix(~ g) # The design matrix

The output after bumphunter:

head(dmrs\$table)
chr     start       end     value     area cluster indexStart indexEnd  L clusterL p.value fwer p.valueArea fwerArea
2685 chr18  74961727  74962672 0.2516403 3.271323   88118     424245   424257 13       35       0    0           0        0

1

The way you specified your model matrix will by default set the second coefficient to be tumor - control. You can test that by looking at your factor levels (the first factor level is by default the baseline level) or by looking at the design matrix itself.

Thank you very much.

Answer: Meaning of the value column in bumphunter output (minfi package)
0
15 months ago by
src30
src30 wrote:

Hi all,

I was wondering how average difference in methylation could be > 1 if beta values range from 0 to 1?

Thank you!