Description Usage Arguments Details Value Author(s) See Also Examples
This function takes the M-values to produce the distance $D$ to be the maximum value satisfies that the proportion of absolute values of two-group difference larger than $D$ is at certain level. Due to the precision limitation, the $D$'s for hyper- and hypomethylation are not necessarily the same. If the samples are not totally paired, than user should first call 'reformData' to process M-values.
1 | boundFinder(mv, prop = 0.1)
|
mv |
The input M-values matrix. If the samples are not totally paired, than user MUST first call "reformData" to process M-values. |
prop |
The proportion that absolute values of two-group difference larger than $D$ must be satisfied. Default value is 0.1 |
The choices of 'prop' are not going to be too extreme or stringent, which will produce to dominated prior. This value should depend on the belief that around 'prop' proportion of loci are differentially methylated. In general 0.1 to 0.2 should be reasonable and and well-performed. Users may also choose different $D$'s for two differentially methylation status freely, in this situation, values around 1.5 to 3 are recommended.
Users must ensure the M-values come from paired samples or has been processed by 'reformData' according to the experiment design.
A two-value vector contains $D_1$ and $D_2$ for the group-mean difference of hypermethylation and hypomethylation respectively. Due to the precision limitation, $D_1$ may not necessarily equal to $D_2$.
Linghao SHEN <sl013@ie.cuhk.edu.hk>
reformData
to tackle unpaired data.
1 2 3 4 5 6 7 8 | # Finding the 5% and 95% quantile of normal samples
set.seed(0)
mv <- cbind(rep(0,100000),rnorm(100000))
boundFinder(mv)
# Output matched the normal p-values
# 5.0% 94.9%
#-1.639578 1.639691
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.