BitSeq-package: Bayesian Inference of Transcripts from Sequencing data

In BitSeq: Transcript expression inference and differential expression analysis for RNA-seq data

Description

The BitSeq package is targeted for transcript expression analysis and differential expression analysis of RNA-seq data in two stage process.

In the first stage it uses Bayesian inference methodology to infer expression of individual transcripts from individual RNA-seq experiments.

The second stage of BitSeq embraces the differential expression analysis of transcript expression. Providing expression estimates from replicates of multiple conditions, Log-Normal model of the estimates is used for inferring the condition mean transcript expression and ranking the transcripts based on the likelihood of differential expression.

Author(s)

Peter Glaus, Antti Honkela and Magnus Rattray Maintainer: Peter Glaus <[e-mail]>

References

Glaus, P., Honkela, A. and Rattray, M. (2012). Identifying differentially expressed transcripts from RNA-seq data with biological variation. Bioinformatics, 28(13), 1721-1728.

Examples

## Not run: 
## basic use
res1 <- getExpression("data-c0b0.sam","ensSelect1.fasta")
res2 <- getExpression("data-c0b1.sam","ensSelect1.fasta")
res3 <- getExpression("data-c1b0.sam","ensSelect1.fasta")
res4 <- getExpression("data-c1b1.sam","ensSelect1.fasta")

deRes <- getDE( list(c(res1$fn, res2$fn), 
                     c(res3$fn, res4$fn)) )
## top 10 differentially expressed
head(deRes$pplr[ order(abs(0.5-deRes$pplr$pplr), decreasing=TRUE ), ], 10)

## advanced use, keeping the intermediate files
parseAlignment( "data-c0b0.sam", 
   outFile = "data-c0b0.prob", 
   trSeqFile = "ensSelect1.fasta",
   trInfoFile = "data.tr",
   uniform = TRUE,
   verbose = TRUE )

estimateExpression( "data-c0b0.prob", 
   outFile = "data-c0b0", 
   outputType = "RPKM",
   trInfoFile = "data.tr",
   MCMC_burnIn = 200,
   MCMC_samplesN = 200,
   MCMC_samplesSave = 100,
   MCMC_scaleReduction = 1.1,
   MCMC_chainsN = 2 )

cond1Files = c("data-c0b0.rpkm","data-c0b1.rpkm")
cond2Files = c("data-c1b1.rpkm","data-c1b1.rpkm")
allConditions = list(cond1Files, cond2Files)

getMeanVariance( allConditions, 
   outFile = "data.means",
   log = TRUE )

estimateHyperPar( allConditions,
   outFile = "data.par",
   meanFile = "data.means",
   verbose = TRUE )

estimateDE( allConditions,
   outFile = "data",
   parFile = "data.par" )


## End(Not run)

BitSeq documentation built on Nov. 8, 2020, 5:25 p.m.