%\VignetteIndexEntry{Differential expression for RNA-seq data with dispersion shrinkage} %\VignettePackage{DSS} \documentclass{article} <>= BiocStyle::latex(use.unsrturl=FALSE) @ \renewcommand{\baselinestretch}{1.3} \usepackage{Sweave} \SweaveOpts{keep.source=TRUE,eps=FALSE,include=TRUE,width=4,height=4} %\newcommand{\Robject}[1]{\texttt{#1}} %\newcommand{\Rpackage}[1]{\textit{#1}} %\newcommand{\Rclass}[1]{\textit{#1}} %\newcommand{\Rfunction}[1]{{\small\texttt{#1}}} \author{Hao Wu \\[1em]Department of Biostatistics and Bioinformatics\\ Emory University\\ Atlanta, GA 303022 \\ [1em] \texttt{hao.wu@emory.edu}} \title{\textsf{\textbf{Differential analyses with DSS}}} \begin{document} \maketitle \tableofcontents %% abstract \begin{abstract} This vignette introduces the use of the Bioconductor package DSS ({\underline D}ispersion {\underline S}hrinkage for {\underline S}equencing data), which is designed for differential analysis based on high-throughput sequencing data. It performs differential expression analyses for RNA-seq, and differential methylation analyses for bisulfite sequencing (BS-seq) data. The core of DSS is a new procedure to estimate and shrink gene- or CpG site-specific dispersions, then conduct Wald tests for differential expression/methylation. Compared with existing methods, DSS provides excellent statistical and computational performance. \end{abstract} \section{Introduction} Recent advances in high-throughput sequencing technology have revolutionized genomic research. For example, RNA-seq is a new technology for measuring the abundance of RNA products. Compared to gene expression microarrays, it provides a better dynamic range and lower signal-to-noise ratio. Bisulfite sequencing (BS-seq) is a new technology for measuring DNA methylation. Compared to capture-based methods such as MeDIP-seq, it provides single-base resolution and eliminates biases associated with CpG density. Fundamental questions for RNA-seq or BS-seq data analyses are whether gene expression regulation or DNA methylation dynamics vary under different biological contexts. Identifying sites or regions exhibiting differential expression (DE) or differential methylation (DM) are thus key tasks in functional genomics research. RNA- or BS-seq experiments typically have a limited number of biological replicates due to cost constraints. This can lead to unstable estimation of within group variance, and subsequently undesirable results from hypothesis testing. Variance shrinkage methods have been widely used in DE analyses based on microarray data. The methods are typically based on a Bayesian hierarchical model, with a prior imposed on the gene-specific variances to provide a basis for information sharing across all genes/CpG sites. In these models, shrinkage is achieved for variance estimation. Using shrunk variance in hypothesis tests has been shown to provide better results. A distinct feature of RNA-seq or BS-seq data is that the measurements are in the form of counts. These data are often assumed to be from the Poisson (for RNA-seq) or Binomial (for BS-seq) distributions. Unlike continuous distributions such as the Gaussian distribution, the variances depend on means in these discrete distributions. This implies that the sample variances do not account for biological variation between replicates, and shrinkage cannot be applied on variances directly. In contrast, we assume that our count data come from the Gamma-Poisson (RNA-seq) or Beta-Binomial (BS-seq) distribution. These distributions can be parameterized by a mean and an over dispersion parameter. The over dispersion parameters, which represent the biological variation for replicates within a treatment group, play a central role in the differential analyses. Here we present a new DE/DM detection algorithm, where shrinkage is performed on the dispersion parameters. We first impose a log-normal prior on the dispersions, and then combine data from all genes/CpG sites to shrink dispersions through a penalized likelihood approach. Finally, we construct Wald tests to test each gene/site for differential expression/methylation. Our results show that the new method provides excellent performance compared to existing methods, especially when the overall dispersion level is high or the number of replicates is small. For details of the hierarchical model, the shrinkage method and test procedure, please read \cite{DE} for differential expression from RNA-seq, and \cite{DML} for differential methylation from BS-seq. %Currently DSS only supports comparison of expression or methylation from %two treatment groups. Methods for more advanced study designs %are under development and will be implemented soon. \section{Using {\tt DSS} for differential expression analysis} \subsection{Single factor experiment} Required inputs for DSS are (1) gene expression values as a matrix of integers, rows are for genes and columns are for samples; and (2) a vector representing experimental designs. The length of the design vector must match the number of columns of input counts. Optionally, normalization factors or additional annotation for genes can be supplied. The basic data container in the package is {\tt SeqCountSet} class, which is directly inherited from {\tt ExpressionSet} class defined in {\tt Biobase}. An object of the class contains all necessary information for a DE analysis: gene expression values, experimental designs, and additional annotations. A typical DE analysis contains the following simple steps. \begin{enumerate} \item Create a {\tt SeqCountSet} object using {\tt newSeqCountSet}. \item Estimate normalization factor using {\tt estNormFactors}. \item Estimate and shrink gene-wise dispersion using {\tt estDispersion} \item Two-group comparison using {\tt waldTest}. \end{enumerate} The usage of DSS is demonstrated in the simple simulation below. \begin{enumerate} \item First load in the library, and make a {\tt SeqCountSet} object from some counts for 2000 genes and 6 samples. <>= library(DSS) counts1=matrix(rnbinom(300, mu=10, size=10), ncol=3) counts2=matrix(rnbinom(300, mu=50, size=10), ncol=3) X1=cbind(counts1, counts2) ## these are 100 DE genes X2=matrix(rnbinom(11400, mu=10, size=10), ncol=6) X=rbind(X1,X2) designs=c(0,0,0,1,1,1) seqData=newSeqCountSet(X, designs) seqData @ \item Estimate normalization factor. <>= seqData=estNormFactors(seqData) @ \item Estimate and shrink gene-wise dispersions <<>>= seqData=estDispersion(seqData) @ \item With the normalization factors and dispersions ready, the two-group comparison can be conducted via a Wald test: <<>>= result=waldTest(seqData, 0, 1) head(result,5) @ \end{enumerate} A higher level wrapper function {\tt DSS.DE} is provided for simple RNA-seq DE analysis in a two-group comparison. User only needs to provide a count matrix and a vector of 0's and 1's representing the design, and get DE test results in one line. A simple example is listed below: <<>>= counts = matrix(rpois(600, 10), ncol=6) designs = c(0,0,0,1,1,1) result = DSS.DE(counts, designs) head(result) @ \subsection{Multifactor experiment} {\tt DSS} provides functionalities for dispersion shrinkage for multifactor experimental designs. Downstream model fitting (through genearlized linear model) and hypothesis testing can be performed using other packages such as {\tt edgeR}, with the dispersions estimated from DSS. Below is an example, based a simple simulation, to illustrate the DE analysis of a crossed design. \begin{enumerate} \item First simulate data for a 2x2 crossed experiments. Note the counts are randomly generated. <>= library(DSS) library(edgeR) counts=matrix(rpois(800, 10), ncol=8) design=data.frame(gender=c(rep("M",4), rep("F",4)), strain=rep(c("WT", "Mutant"),4)) X=model.matrix(~gender+strain, data=design) @ \item make SeqCountSet, then estimate size factors and dispersion <>= seqData=newSeqCountSet(counts, as.data.frame(X)) seqData=estNormFactors(seqData) seqData=estDispersion(seqData) @ \item Using edgeR's function to do glm model fitting, but plugging in the estimated size factors and dispersion from DSS. <<>>= fit.edgeR <- glmFit(counts, X, lib.size=normalizationFactor(seqData), dispersion=dispersion(seqData)) @ \item Using edgeR's function to do hypothesis testing on the second parameter of the model (gender). <>= lrt.edgeR <- glmLRT(glmfit=fit.edgeR, coef=2) head(lrt.edgeR$table) @ \end{enumerate} %%% DML detection \section{Using {\tt DSS} for differential methylation analysis} \subsection{Overview} To detect differential methylation, statistical tests are conducted at each CpG site, and then the differential methylation loci (DML) or differential methylation regions (DMR) are called based on user specified threshold. A rigorous statistical tests should account for biological variations among replicates and the coverage depth. Most existing methods for DM analysis are based on {\it ad hoc} methods. For example, using Fisher's exact ignores the biological variations, using t-test on estimated methylation levels ignores the coverage depth. Sometimes arbitrary filtering are recommended: loci with coverages lower than an arbitrary threshold are filtered out. The DM detection procedure implemented in DSS is based on a rigorous Wald test for beta-binomial distributions. The test statistics depend on the biological variations (captured by dispersion parameter) as well as the coverage depth. An important part of the algorithm is the estimation of dispersion parameter, which is achieved through a shrinkage estimator based on a Bayesian hierarchical model \cite{DML}. A great advantage of DSS is that the test can be performed even when there is no biological replicates. That's because by smoothing procedure, the neighboring CpG sites can be viewed as ``pseudo-replicates", and the dispersion can still be estimated with reasonable precision. This package depends on {\tt bsseq} Bioconductor package, which has neat definition of data structures and many useful utility functions. In order to use the DM detection functionalities, {\tt bsseq} needs to be pre-installed. \subsection{Input data} For one BS-seq experiments, after sequence alignment and some processing, the BS-seq data are usually summarized into following information for each CG position: chromosome number, genomic coordinate, total number of reads, and number of reads showing methylation. For a sample, this information are saved in a simple text file, with each row representing a CpG site. Below shows an example of a small part of such a file: \begin{verbatim} chr pos N X chr18 3014904 26 2 chr18 3031032 33 12 chr18 3031044 33 13 chr18 3031065 48 24 chr18 3031069 17 4 chr18 3031082 93 37 chr18 3031089 76 25 chr18 3031092 76 28 \end{verbatim} DML/DMR detection using {\tt DSS} starts from several such text files. A typical DML detection contains two simple steps. First one conduct DM test at each CpG site, then DML/DMR are called based on the test result and user specified threshold. We will use example data distributed with {\tt DSS} to illustrate the procedure. \subsection{DML/DMR detection from two-group comparison} Below are the steps to call DML or DMR for BS-seq data in two-group comparison setting. \begin{enumerate} \item Load in library. Read in text files and create an object of {\tt BSseq} class, which is defined in {\tt bsseq} Bioconductor package. This step requires {\tt bsseq} Bioconductor package. {\tt BSseq} class is defined in that package. <<>>= library(DSS) require(bsseq) path <- file.path(system.file(package="DSS"), "extdata") dat1.1 <- read.table(file.path(path, "cond1_1.txt"), header=TRUE) dat1.2 <- read.table(file.path(path, "cond1_2.txt"), header=TRUE) dat2.1 <- read.table(file.path(path, "cond2_1.txt"), header=TRUE) dat2.2 <- read.table(file.path(path, "cond2_2.txt"), header=TRUE) BSobj <- makeBSseqData( list(dat1.1, dat1.2, dat2.1, dat2.2), c("C1","C2", "N1", "N2") )[1:10000,] BSobj @ \item Perform statistical test for DML by calling {\tt DMLtest} function. This function basically performs following steps: (1) estimate mean methylation levels for all CpG site; (2) estimate dispersions at each CpG sites; (3) conduct Wald test. For the first step, there's an option for smoothing or not. Because the methylation levels show strong spatial correlations, smoothing can help obtain better estimates of mean methylation when the CpG sites are dense in the data (such as from the whole-genome BS-seq). However for data with sparse CpG, such as from RRBS or hydroxyl-methylation, smoothing is not recommended. To perform DML test without smoothing, do: <<>>= dmlTest <- DMLtest(BSobj, group1=c("C1", "C2"), group2=c("N1", "N2")) head(dmlTest) @ To perform statistical test for DML with smoothing, do: <<>>= dmlTest.sm <- DMLtest(BSobj, group1=c("C1", "C2"), group2=c("N1", "N2"), smoothing=TRUE) @ There are two options for smoothing: a simple moving average, or the BSmooth method implemented in {\tt bsseq} package. The BSmooth method produces much smoother curve, which is good for visualization purpose. However, it is very computationally intensive, and the results are not very different from moving average in terms of DMR calling. So we recommend using moving average. Smoothing span is an important parameter in smoothing procedure and have non-trivial impact on DMR calling. We use 500 bp as default, and think that it performs well in real data tests. \item With the test results, one can call DML by using {\tt callDML} function. The results DMLs are sorted by the significance. <<>>= dmls <- callDML(dmlTest, p.threshold=0.001) head(dmls) @ By default, the test is based on the null hypothesis that the difference in methylation levels is 0. Alternatively, users can specify a threshold for difference. For example, to detect loci with difference greater than 0.1, do: <<>>= dmls2 <- callDML(dmlTest, delta=0.1, p.threshold=0.001) head(dmls2) @ When delta is specified, the function will compute the posterior probability that the difference of the means is greater than delta. So technically speaking, the threshold for p-value here actually refers to the threshold for 1-posterior probability, or the local FDR. Here we use the same parameter name for the sake of the consistence of function syntax. \item DMR detection is also Based on the DML test results, by calling {\tt callDMR} function. Regions with many statistically significant CpG sites are identified as DMRs. Some restrictions are provided by users, including the minimum length, minimum number of CpG sites, percentage of CpG site being significant in the region, etc. There are some {\it post hoc} procedures to merge nearby DMRs into longer ones. <<>>= dmrs <- callDMR(dmlTest, p.threshold=0.01) head(dmrs) @ Here the DMRs are sorted by ``areaStat", which is defined in {\tt bsseq} as the sum of the test statistics of all CpG sites within the DMR. Similarly, users can specify a threshold for difference. For example, to detect regions with difference greater than 0.1, do: <<>>= dmrs2 <- callDMR(dmlTest, delta=0.1, p.threshold=0.05) head(dmrs2) @ Note that the distribution of test statistics (and p-values) depends on the differences in methylation levels and biological variations, as well as technical factors such as coverage depth. It is very difficulty to select a natural and rigorous threshold for defining DMRs. We recommend users try different thresholds in order to obtain satisfactory results. \item The DMRs can be visualized using {\tt showOneDMR} function, This function provides more information than the {\tt plotRegion} function in {\tt bsseq}. It plots the methylation percentages as well as the coverage depths at each CpG sites, instead of just the smoothed curve. So the coverage depth information will be available in the figure. To use the function, do <>= showOneDMR(dmrs[1,], BSobj) @ The result figure looks like the following. Note that the figure below is not generated from the above example. The example data are from RRBS experiment so the DMRs are much shorter. \begin{figure}[h!] \centerline{\includegraphics[width=5.5in]{aDMR.pdf}} \end{figure} \end{enumerate} \newpage \subsection{DML/DMR detection from general experimental design} In DSS, BS-seq data from a general experimental design (such as crossed experiment, or experiment with covariates) is modeled through a generalized linear model framework. We use arcsine link function instead of the typical logit link for it better deals with data at boundaries (methylation levels close to 0 or 1). Linear model fitting is done through ordinary least square on transformed methylation levels. Standard errors for the estimates are derived with consideration of count data distribution and transformation. A Wald test is applied to perform hypothesis testing. \begin{enumerate} \item Load in data distributed with {\tt DSS}. This is a small portion of a set of RRBS experiments. There are 5000 CpG sites and 16 samples. The experiment is a $2\time 2$ design (2 cases and 2 cell types). There are 4 replicates in each case:cell combination. <<>>= data(RRBS) RRBS design @ \item Fit a linear model using {\tt DMLfit.multiFactor} function, include case, cell, and case:cell interaction. <<>>= DMLfit = DMLfit.multiFactor(RRBS, design=design, formula=~case+cell+case:cell) @ \item Use {\tt DMLtest.multiFactor} function to test the cell effect. It is important to note that the {\tt coef} parameter is the index of the coefficient to be tested for being 0. Because the model (as specified by {\tt formula} in {\tt DMLfit.multiFactor}) include intercept, the cell effect is the 3rd column in the design matrix, so we use {\tt coef=3} here. <<>>= DMLtest.cell = DMLtest.multiFactor(DMLfit, coef=3) @ Result from this step is a data frame with chromosome number, CpG site position, test statistics, p-values (from normal distribution), and FDR. Rows are sorted by chromosome/position of the CpG sites. To obtain top ranked CpG sites, one can sort the data frame using following codes: <<>>= ix=sort(DMLtest.cell[,"pvals"], index.return=TRUE)$ix head(DMLtest.cell[ix,]) @ Below is a figure showing the distributions of test statistics and p-values from this example dataset <>= par(mfrow=c(1,2)) hist(DMLtest.cell$stat, 50, main="test statistics", xlab="") hist(DMLtest.cell$pvals, 50, main="P values", xlab="") @ \end{enumerate} These procedures are computationally very efficient. For a typical RRBS dataset with 4 million CpG sites, it takes around 20 minutes. In comparison, other methods such as RADMeth or BiSeq takes at least 10 times longer. \newpage \section{Session Info} <>= sessionInfo() @ \begin{thebibliography}{99} \bibitem{DML} \textsc{Hao Feng, Karen Conneely and Hao Wu}. (2014). \newblock A bayesian hierarchical model to detect differentially methylated loci from single nucleotide resolution sequencing data. \newblock {\em Nucleic acids research\/}~\textbf{42}(8), e69--e69. \bibitem{DMR1} \textsc{Hao Wu, Tianlei Xu, Hao Feng, Li Chen, Ben Li, Bing Yao, Zhaohui Qin, Peng Jin and Karen N. Conneely}. (2015). \newblock Detection of differentially methylated regions from whole-genome bisulfite sequencing data without replicates. \newblock {\em Nucleic acids research\/}~doi: 10.1093/nar/gkv715. \bibitem{DE} \textsc{Hao Wu, Chi Wang and Zhijing Wu}. (2013). \newblock A new shrinkage estimator for dispersion improves differential expression detection in rna-seq data. \newblock {\em Biostatistics\/}~\textbf{14}(2), 232--243. \end{thebibliography} \end{document}