Question

bumphunter output clarification

0

Entering edit mode

chelsey.ju • 0

@chelseyju-12100

Last seen 7.0 years ago

Hello,

I am looking for clarification and explanations for the following output headers of bumphunter DMR analysis:

area

cluster

indexStart

indexEnd

L

cluster

p,value

p,valueArea

Additionally, if a B=1000 permutation cut off was used, do the p values need to be further adjusted for multiple testing analysis or are they already corrected?

thanks!

bumphunter DMR DMR analysis minfi bioconductor • 4.1k views

ADD COMMENT • link updated 7.3 years ago by rafa ▴ 60 • written 7.3 years ago by chelsey.ju • 0

score 1 · Answer 1 · 2017-01-04

1

Entering edit mode

rafa ▴ 60

@rafa-7160

Last seen 5.5 years ago

United States

I would start by reading the Details section of the bumphunter help file.

Regarding the table, each row is a candidate DMR.

value - is the average of the estimated coefficient (usually representing the difference between two groups) in the region

area - is the absolute value of the sum of the estimated coefficients in the region

cluster - the genome is divided into smaller regions, called clusters. We search for regions in each cluster to avoid going over a long stretch of the genome with no data. The value provided here is the ID for the cluster.

indexStart - is the index of the probe in the original methylation matrix where the region starts

indexEnd- is the index of the probe in the original methylation matrix where the region ends

L - is the number of probes in the region

clusterL - number of probes in the cluster

p.value - are explained in the help file. These do need multiple comparison corrections.

fwer - are explained in the help file. These do not need multiple comparison corrections

ADD COMMENT • link 7.3 years ago rafa ▴ 60

0

Entering edit mode

This is helpful! Can anyone clarify what the difference is between the region and the cluster on this bit?:

L - is the number of probes in the region

clusterL - number of probes in the cluster

ADD REPLY • link 7.3 years ago brhead ▴ 40

0

Entering edit mode

A cluster is defined purely based on the array design (the spacing between probes). It is a collection of probes where the between probe distance (on the genome) is smaller than X. If X is set to be Inf, each chromosome will be a cluster. A region is (always) a subset of a cluster and is a set of probes which are differentially methylated. So that depends heavily on the actual observed data. Best, Kasper On Fri, Jan 6, 2017 at 3:12 AM, brhead [bioc] <noreply@bioconductor.org> wrote: > Activity on a post you are following on support.bioconductor.org > > User brhead <https: support.bioconductor.org="" u="" 11927=""/> wrote Comment: > bumphunter output clarification > <https: support.bioconductor.org="" p="" 90751="" #90845="">: > > This is helpful! Can anyone clarify what the difference is between the > region and the cluster on this bit?: > > L - is the number of probes in the region > > clusterL - number of probes in the cluster > > > > ------------------------------ > > Post tags: bumphunter, DMR, DMR analysis, minfi, bioconductor > > You may reply via email or visit https://support.bioconductor. > org/p/90751/#90845 >