Methylphet-package: Base-resolution methylation patterns accurately predict...

In benliemory/Methylphet: Base-resolution methylation patterns accurately predict transcription factor bindings in vivo

Description

Methylphet adopts a two-step method to predict transcription factor-DNA interaction using DNA methylation profiles from whole-genome bisulfite sequencing data by exploiting the connection between DNA methylation level and transcription factor binding. In the first step, beta-binomial models are devised to characterize DNA methylation data around TF binding sites and the background to estimate methylation scores. Along with other static genomic features, a random forest framework is adopted in the second step to predict transcription factor-DNA interaction. When all methylation profile are taken together and combined with features at the sequence level, Methylphet can accurately predict TF binding and performs favorably when compared against competing methods.

Details

Package:	Methylphet
Type:	Package
Version:	1.0
Date:	2014-08-21
License:	GPL-2

Train data set and test data set are required for the prediction of TFBS. Both data set requires location informatoin for candidate sites, motif scores, methylation information(CpG or CpH) or 5hmc information. Training data set needs additional golden standard or ChIP-seq information. With all these information provided, Methylphet will return predict results for each candidate site as well as the predict model.

Author(s)

Tianlei Xu, Ben Li, Hao Wu, Zhaohui Qin

Maintainer: Tianlei Xu<[e-mail]>, Ben Li<[e-mail]>, Hao Wu<[e-mail]>, Zhaohui Qin<[e-mail]>

Examples

### Load the package
library("Methylphet")


### Load all the data needed 
data(list = data(package="Methylphet")$results[,3])

### Using CpG methylation information only to predict TFBS when 0/1 golden standard is provided.
OCT4_CpG = Methylphet(traindata.mat = mESdata.motif.chr10,traindata.methyl1=mESdata.CpG.chr10, 
                      goldstandard=goldstandard.chr10, OtherGenomicFeatures.train=OtherGenomicFeatures.mES.chr10,
                      testdata.mat =H1data.motif.chr10, testdata.methyl1=H1data.CpG.chr10,
                      OtherGenomicFeatures.test=OtherGenomicFeatures.H1.chr10)

### Using both CpG and CpH methylation information to predict TFBS when location for ChIP-seq peaks are provided.
OCT4_CpG_CpH = Methylphet(traindata.mat = mESdata.motif.chr10,
                       traindata.methyl1=mESdata.CpG.chr10,traindata.methyl2=mESdata.CpH.chr10,
                          ChIPseqPeaks = peak.GR.ES.chr10, OtherGenomicFeatures.train=OtherGenomicFeatures.mES.chr10,
                      testdata.mat =H1data.motif.chr10, 
                          testdata.methyl1=H1data.CpG.chr10,testdata.methyl2=H1data.CpH.chr10,
                      OtherGenomicFeatures.test=OtherGenomicFeatures.H1.chr10)

benliemory/Methylphet documentation built on May 12, 2019, 12:16 p.m.