Contents

Note: the most recent version of this tutorial can be found here and a short overview slide show here.

1 Introduction

systemPipeR provides utilities for building and running automated end-to-end analysis workflows for a wide range of next generation sequence (NGS) applications such as RNA-Seq, ChIP-Seq, VAR-Seq and Ribo-Seq (Girke 2014). Important features include a uniform workflow interface across different NGS applications, automated report generation, and support for running both R and command-line software, such as NGS aligners or peak/variant callers, on local computers or compute clusters. The latter supports interactive job submissions and batch submissions to queuing systems of clusters. For instance, systemPipeR can be used with most command-line aligners such as BWA (Heng Li 2013; H Li and Durbin 2009), TopHat2 (Kim et al. 2013) and Bowtie2 (Langmead and Salzberg 2012), as well as the R-based NGS aligners Rsubread (Liao, Smyth, and Shi 2013) and gsnap (gmapR) (Wu and Nacu 2010). Efficient handling of complex sample sets (e.g. FASTQ/BAM files) and experimental designs is facilitated by a well-defined sample annotation infrastructure which improves reproducibility and user-friendliness of many typical analysis workflows in the NGS area (Lawrence et al. 2013).

Motivation and advantages of sytemPipeR environment:

  1. Facilitates design of complex NGS workflows involving multiple R/Bioconductor packages
  2. Common workflow interface for different NGS applications
  3. Makes NGS analysis with Bioconductor utilities more accessible to new users
  4. Simplifies usage of command-line software from within R
  5. Reduces complexity of using compute clusters for R and command-line software
  6. Accelerates runtime of workflows via parallelzation on computer systems with mutiple CPU cores and/or multiple compute nodes
  7. Automates generation of analysis reports to improve reproducibility

A central concept for designing workflows within the sytemPipeR environment is the use of workflow management containers called SYSargs (see Figure 1). Instances of this S4 object class are constructed by the systemArgs function from two simple tabular files: a targets file and a param file. The latter is optional for workflow steps lacking command-line software. Typically, a SYSargs instance stores all sample-level inputs as well as the paths to the corresponding outputs generated by command-line- or R-based software generating sample-level output files, such as read preprocessors (trimmed/filtered FASTQ files), aligners (SAM/BAM files), variant callers (VCF/BCF files) or peak callers (BED/WIG files). Each sample level input/outfile operation uses its own SYSargs instance. The outpaths of SYSargs usually define the sample inputs for the next SYSargs instance. This connectivity is established by writing the outpaths with the writeTargetsout function to a new targets file that serves as input to the next systemArgs call. Typically, the user has to provide only the initial targets file. All downstream targets files are generated automatically. By chaining several SYSargs steps together one can construct complex workflows involving many sample-level input/output file operations with any combinaton of command-line or R-based software.

Figure 1: Workflow design structure of systemPipeR


The intended way of running sytemPipeR workflows is via *.Rnw or *.Rmd files, which can be executed either line-wise in interactive mode or with a single command from R or the command-line using a Makefile. This way comprehensive and reproducible analysis reports can be generated in PDF or HTML format in a fully automated manner by making use of the highly functional reporting utilities available for R. Templates for setting up custom project reports are provided as *.Rnw files by the helper package systemPipeRdata and in the vignettes subdirectory of systemPipeR. The corresponding PDFs of these report templates are available here: systemPipeRNAseq, systemPipeRIBOseq, systemPipeChIPseq and systemPipeVARseq. To work with *.Rnw or *.Rmd files efficiently, basic knowledge of Sweave or knitr and Latex or R Markdown v2 is required.

Back to Table of Contents

2 Getting Started

2.1 Installation

The R software for running systemPipeR can be downloaded from CRAN. The systemPipeR environment can be installed from the R console using the biocLite install command. The associated data package systemPipeRdata can be installed the same way. The latter is a helper package for generating systemPipeR workflow environments with a single command containing all parameter files and sample data required to quickly test and run workflows.

source("http://bioconductor.org/biocLite.R") # Sources the biocLite.R installation script 
biocLite("systemPipeR") # Installs systemPipeR 
biocLite("systemPipeRdata") # Installs systemPipeRdata
Back to Table of Contents

2.2 Loading package and documentation

library("systemPipeR") # Loads the package
library(help="systemPipeR") # Lists package info
vignette("systemPipeR") # Opens vignette
Back to Table of Contents

2.3 Load sample data and workflow templates

The mini sample FASTQ files used by this overview vignette as well as the associated workflow reporting vignettes can be loaded via the systemPipeRdata package as shown below. The chosen data set SRP010938 contains 18 paired-end (PE) read sets from Arabidposis thaliana (Howard et al. 2013). To minimize processing time during testing, each FASTQ file has been subsetted to 90,000-100,000 randomly sampled PE reads that map to the first 100,000 nucleotides of each chromosome of the A. thalina genome. The corresponding reference genome sequence (FASTA) and its GFF annotion files (provided in the same download) have been truncated accordingly. This way the entire test sample data set requires less than 200MB disk storage space. A PE read set has been chosen for this test data set for flexibility, because it can be used for testing both types of analysis routines requiring either SE (single end) reads or PE reads.

The following generates a fully populated systemPipeR workflow environment (here for RNA-Seq) in the current working directory of an R session. At this time the package includes workflow templates for RNA-Seq, ChIP-Seq, VAR-Seq and Ribo-Seq. Templates for additional NGS applications will be provided in the future.

library(systemPipeRdata)
genWorkenvir(workflow="rnaseq")
setwd("rnaseq")

The working environment of the sample data loaded in the previous step contains the following preconfigured directory structure. Directory names are indicated in grey. Users can change this structure as needed, but need to adjust the code in their workflows accordingly.

The following parameter files are included in each workflow template:

  1. targets.txt: initial one provided by user; downstream targets_*.txt files are generated automatically
  2. *.param: defines parameter for input/output file operations, e.g. trim.param, bwa.param, vartools.parm, …
  3. *_run.sh: optional bash script, e.g.: gatk_run.sh
  4. Compute cluster environment (skip on single machine):
    • .BatchJobs: defines type of scheduler for BatchJobs
    • *.tmpl: specifies parameters of scheduler used by a system, e.g. Torque, SGE, StarCluster, Slurm, etc.
Back to Table of Contents

2.4 Structure of targets file

The targets file defines all input files (e.g. FASTQ, BAM, BCF) and sample comparisons of an analysis workflow. The following shows the format of a sample targets file included in the package. It also can be viewed and downloaded from systemPipeR’s GitHub repository here. In a target file with a single type of input files, here FASTQ files of single end (SE) reads, the first three columns are mandatory including their column names, while it is four mandatory columns for FASTQ files of PE reads. All subsequent columns are optional and any number of additional columns can be added as needed.

2.4.1 Structure of targets file for single end (SE) samples

library(systemPipeR)
targetspath <- system.file("extdata", "targets.txt", package="systemPipeR") 
read.delim(targetspath, comment.char = "#")
##                    FileName SampleName Factor SampleLong Experiment        Date
## 1  ./data/SRR446027_1.fastq        M1A     M1  Mock.1h.A          1 23-Mar-2012
## 2  ./data/SRR446028_1.fastq        M1B     M1  Mock.1h.B          1 23-Mar-2012
## 3  ./data/SRR446029_1.fastq        A1A     A1   Avr.1h.A          1 23-Mar-2012
## 4  ./data/SRR446030_1.fastq        A1B     A1   Avr.1h.B          1 23-Mar-2012
## 5  ./data/SRR446031_1.fastq        V1A     V1   Vir.1h.A          1 23-Mar-2012
## 6  ./data/SRR446032_1.fastq        V1B     V1   Vir.1h.B          1 23-Mar-2012
## 7  ./data/SRR446033_1.fastq        M6A     M6  Mock.6h.A          1 23-Mar-2012
## 8  ./data/SRR446034_1.fastq        M6B     M6  Mock.6h.B          1 23-Mar-2012
## 9  ./data/SRR446035_1.fastq        A6A     A6   Avr.6h.A          1 23-Mar-2012
## 10 ./data/SRR446036_1.fastq        A6B     A6   Avr.6h.B          1 23-Mar-2012
## 11 ./data/SRR446037_1.fastq        V6A     V6   Vir.6h.A          1 23-Mar-2012
## 12 ./data/SRR446038_1.fastq        V6B     V6   Vir.6h.B          1 23-Mar-2012
## 13 ./data/SRR446039_1.fastq       M12A    M12 Mock.12h.A          1 23-Mar-2012
## 14 ./data/SRR446040_1.fastq       M12B    M12 Mock.12h.B          1 23-Mar-2012
## 15 ./data/SRR446041_1.fastq       A12A    A12  Avr.12h.A          1 23-Mar-2012
## 16 ./data/SRR446042_1.fastq       A12B    A12  Avr.12h.B          1 23-Mar-2012
## 17 ./data/SRR446043_1.fastq       V12A    V12  Vir.12h.A          1 23-Mar-2012
## 18 ./data/SRR446044_1.fastq       V12B    V12  Vir.12h.B          1 23-Mar-2012

To work with custom data, users need to generate a targets file containing the paths to their own FASTQ files and then provide under targetspath the path to the corresponding targets file.

Back to Table of Contents

2.4.2 Structure of targets file for paired end (PE) samples

targetspath <- system.file("extdata", "targetsPE.txt", package="systemPipeR")
read.delim(targetspath, comment.char = "#")[1:2,1:6]
##                  FileName1                FileName2 SampleName Factor SampleLong Experiment
## 1 ./data/SRR446027_1.fastq ./data/SRR446027_2.fastq        M1A     M1  Mock.1h.A          1
## 2 ./data/SRR446028_1.fastq ./data/SRR446028_2.fastq        M1B     M1  Mock.1h.B          1
Back to Table of Contents

2.4.3 Sample comparisons

Sample comparisons are defined in the header lines of the targets file starting with ‘# <CMP>’.

readLines(targetspath)[1:4]
## [1] "# Project ID: Arabidopsis - Pseudomonas alternative splicing study (SRA: SRP010938; PMID: 24098335)"                                                                              
## [2] "# The following line(s) allow to specify the contrasts needed for comparative analyses, such as DEG identification. All possible comparisons can be specified with 'CMPset: ALL'."
## [3] "# <CMP> CMPset1: M1-A1, M1-V1, A1-V1, M6-A6, M6-V6, A6-V6, M12-A12, M12-V12, A12-V12"                                                                                             
## [4] "# <CMP> CMPset2: ALL"

The function readComp imports the comparison information and stores it in a list. Alternatively, readComp can obtain the comparison information from the corresponding SYSargs object (see below). Note, these header lines are optional. They are mainly useful for controlling comparative analyses according to certain biological expectations, such as identifying differentially expressed genes in RNA-Seq experiments based on simple pair-wise comparisons.

readComp(file=targetspath, format="vector", delim="-")
## $CMPset1
## [1] "M1-A1"   "M1-V1"   "A1-V1"   "M6-A6"   "M6-V6"   "A6-V6"   "M12-A12" "M12-V12" "A12-V12"
## 
## $CMPset2
##  [1] "M1-A1"   "M1-V1"   "M1-M6"   "M1-A6"   "M1-V6"   "M1-M12"  "M1-A12"  "M1-V12"  "A1-V1"  
## [10] "A1-M6"   "A1-A6"   "A1-V6"   "A1-M12"  "A1-A12"  "A1-V12"  "V1-M6"   "V1-A6"   "V1-V6"  
## [19] "V1-M12"  "V1-A12"  "V1-V12"  "M6-A6"   "M6-V6"   "M6-M12"  "M6-A12"  "M6-V12"  "A6-V6"  
## [28] "A6-M12"  "A6-A12"  "A6-V12"  "V6-M12"  "V6-A12"  "V6-V12"  "M12-A12" "M12-V12" "A12-V12"
Back to Table of Contents

2.5 Structure of param file and SYSargs container

The param file defines the parameters of a chosen command-line software. The following shows the format of a sample param file provided by this package.

parampath <- system.file("extdata", "tophat.param", package="systemPipeR")
read.delim(parampath, comment.char = "#")
##      PairSet         Name                                  Value
## 1    modules         <NA>                          bowtie2/2.2.5
## 2    modules         <NA>                          tophat/2.0.14
## 3   software         <NA>                                 tophat
## 4      cores           -p                                      4
## 5      other         <NA> -g 1 --segment-length 25 -i 30 -I 3000
## 6   outfile1           -o                            <FileName1>
## 7   outfile1         path                             ./results/
## 8   outfile1       remove                                   <NA>
## 9   outfile1       append                                .tophat
## 10  outfile1 outextension              .tophat/accepted_hits.bam
## 11 reference         <NA>                    ./data/tair10.fasta
## 12   infile1         <NA>                            <FileName1>
## 13   infile1         path                                   <NA>
## 14   infile2         <NA>                            <FileName2>
## 15   infile2         path                                   <NA>

The systemArgs function imports the definitions of both the param file and the targets file, and stores all relevant information in a SYSargs object (S4 class). To run the pipeline without command-line software, one can assign NULL to sysma instead of a param file. In addition, one can start systemPipeR workflows with pre-generated BAM files by providing a targets file where the FileName column provides the paths to the BAM files. Note, in the following example the usage of suppressWarnings() is only relevant for building this vignette. In typical workflows it should be removed.

args <- suppressWarnings(systemArgs(sysma=parampath, mytargets=targetspath))
args
## An instance of 'SYSargs' for running 'tophat' on 18 samples

Several accessor methods are available that are named after the slot names of the SYSargs object.

names(args)
##  [1] "targetsin"     "targetsout"    "targetsheader" "modules"       "software"      "cores"        
##  [7] "other"         "reference"     "results"       "infile1"       "infile2"       "outfile1"     
## [13] "sysargs"       "outpaths"

Of particular interest is the sysargs() method. It constructs the system commands for running command-lined software as specified by a given param file combined with the paths to the input samples (e.g. FASTQ files) provided by a targets file. The example below shows the sysargs() output for running TopHat2 on the first PE read sample. Evaluating the output of sysargs() can be very helpful for designing and debugging param files of new command-line software or changing the parameter settings of existing ones.

sysargs(args)[1]
##                                                                                                                                                                                                                                                                          M1A 
## "tophat -p 4 -g 1 --segment-length 25 -i 30 -I 3000 -o /tmp/RtmpUYzJqX/Rbuild200f76278aa/systemPipeR/vignettes/results/SRR446027_1.fastq.tophat /tmp/RtmpUYzJqX/Rbuild200f76278aa/systemPipeR/vignettes/data/tair10.fasta ./data/SRR446027_1.fastq ./data/SRR446027_2.fastq"
modules(args)
## [1] "bowtie2/2.2.5" "tophat/2.0.14"
cores(args)
## [1] 4
outpaths(args)[1]
##                                                                                                          M1A 
## "/tmp/RtmpUYzJqX/Rbuild200f76278aa/systemPipeR/vignettes/results/SRR446027_1.fastq.tophat/accepted_hits.bam"

The content of the param file can also be returned as JSON object as follows (requires rjson package).

systemArgs(sysma=parampath, mytargets=targetspath, type="json")
## [1] "{\"modules\":{\"n1\":\"\",\"v2\":\"bowtie2/2.2.5\",\"n1\":\"\",\"v2\":\"tophat/2.0.14\"},\"software\":{\"n1\":\"\",\"v1\":\"tophat\"},\"cores\":{\"n1\":\"-p\",\"v1\":\"4\"},\"other\":{\"n1\":\"\",\"v1\":\"-g 1 --segment-length 25 -i 30 -I 3000\"},\"outfile1\":{\"n1\":\"-o\",\"v2\":\"<FileName1>\",\"n3\":\"path\",\"v4\":\"./results/\",\"n5\":\"remove\",\"v1\":\"\",\"n2\":\"append\",\"v3\":\".tophat\",\"n4\":\"outextension\",\"v5\":\".tophat/accepted_hits.bam\"},\"reference\":{\"n1\":\"\",\"v1\":\"./data/tair10.fasta\"},\"infile1\":{\"n1\":\"\",\"v2\":\"<FileName1>\",\"n1\":\"path\",\"v2\":\"\"},\"infile2\":{\"n1\":\"\",\"v2\":\"<FileName2>\",\"n1\":\"path\",\"v2\":\"\"}}"
Back to Table of Contents

3 Workflow overview

3.1 Define environment settings and samples

A typical workflow starts with generating the expected working environment containing the proper directory structure, input files and parameter settings. To simplify this task, one can load one of the existing NGS workflows templates provided by systemPipeRdata into the current working directory. The following does this for the rnaseq template. The name of the resulting workflow directory can be specified under the mydirname argument. The default NULL uses the name of the chosen workflow. An error is issued if a directory of the same name and path exists already. On Linux and OS X systems one can also create new workflow instances from the command-line of a terminal as shown here. To apply workflows to custom data, the user needs to modify the targets file and if necessary update the corresponding param file(s). A collection of pre-generated param files is provided in the param subdirectory of each workflow template. They are also viewable in the GitHub repository of systemPipeRdata (see here).

library(systemPipeR)
library(systemPipeRdata)
genWorkenvir(workflow="rnaseq", mydirname=NULL)
setwd("rnaseq")

Construct SYSargs object from param and targets files.

args <- systemArgs(sysma="param/trim.param", mytargets="targets.txt")
Back to Table of Contents

3.2 Read Preprocessing

The function preprocessReads allows to apply predefined or custom read preprocessing functions to all FASTQ files referenced in a SYSargs container, such as quality filtering or adaptor trimming routines. The paths to the resulting output FASTQ files are stored in the outpaths slot of the SYSargs object. Internally, preprocessReads uses the FastqStreamer function from the ShortRead package to stream through large FASTQ files in a memory-efficient manner. The following example performs adaptor trimming with the trimLRPatterns function from the Biostrings package. After the trimming step a new targets file is generated (here targets_trim.txt) containing the paths to the trimmed FASTQ files. The new targets file can be used for the next workflow step with an updated SYSargs instance, e.g. running the NGS alignments with the trimmed FASTQ files.

preprocessReads(args=args, Fct="trimLRPatterns(Rpattern='GCCCGGGTAA', subject=fq)", 
                batchsize=100000, overwrite=TRUE, compress=TRUE)
writeTargetsout(x=args, file="targets_trim.txt")

The following example shows how one can design a custom read preprocessing function using utilities provided by the ShortRead package, and then run it in batch mode with the ‘preprocessReads’ function (here on paired-end reads).

args <- systemArgs(sysma="param/trimPE.param", mytargets="targetsPE.txt")
filterFct <- function(fq, cutoff=20, Nexceptions=0) {
    qcount <- rowSums(as(quality(fq), "matrix") <= cutoff)
    fq[qcount <= Nexceptions] # Retains reads where Phred scores are >= cutoff with N exceptions
}
preprocessReads(args=args, Fct="filterFct(fq, cutoff=20, Nexceptions=0)", batchsize=100000)
writeTargetsout(x=args, file="targets_PEtrim.txt")
Back to Table of Contents

3.3 FASTQ quality report

The following seeFastq and seeFastqPlot functions generate and plot a series of useful quality statistics for a set of FASTQ files including per cycle quality box plots, base proportions, base-level quality trends, relative k-mer diversity, length and occurrence distribution of reads, number of reads above quality cutoffs and mean quality distribution.
The function seeFastq computes the quality statistics and stores the results in a relatively small list object that can be saved to disk with save() and reloaded with load() for later plotting. The argument klength specifies the k-mer length and batchsize the number of reads to random sample from each FASTQ file.

fqlist <- seeFastq(fastq=infile1(args), batchsize=10000, klength=8)
pdf("./results/fastqReport.pdf", height=18, width=4*length(fqlist))
seeFastqPlot(fqlist)
dev.off()
Figure 2: FASTQ quality report


Back to Table of Contents

Parallelization of QC report on single machine with multiple cores

args <- systemArgs(sysma="param/tophat.param", mytargets="targets.txt")
f <- function(x) seeFastq(fastq=infile1(args)[x], batchsize=100000, klength=8)
fqlist <- bplapply(seq(along=args), f, BPPARAM = MulticoreParam(workers=8))
seeFastqPlot(unlist(fqlist, recursive=FALSE))
Back to Table of Contents

Parallelization of QC report via scheduler (e.g. Torque) across several compute nodes

library(BiocParallel); library(BatchJobs)
f <- function(x) {
    library(systemPipeR)
    args <- systemArgs(sysma="param/tophat.param", mytargets="targets.txt")
    seeFastq(fastq=infile1(args)[x], batchsize=100000, klength=8)
}
funs <- makeClusterFunctionsTorque("torque.tmpl")
param <- BatchJobsParam(length(args), resources=list(walltime="20:00:00", nodes="1:ppn=1", memory="6gb"), cluster.functions=funs)
register(param)
fqlist <- bplapply(seq(along=args), f)
seeFastqPlot(unlist(fqlist, recursive=FALSE))
Back to Table of Contents

3.4 Alignment with Tophat2

Build Bowtie2 index.

args <- systemArgs(sysma="param/tophat.param", mytargets="targets.txt")
moduleload(modules(args)) # Skip if module system is not available
system("bowtie2-build ./data/tair10.fasta ./data/tair10.fasta")

Execute SYSargs on a single machine without submitting to a queuing system of a compute cluster. This way the input FASTQ files will be processed sequentially. If available, multiple CPU cores can be used for processing each file. The number of CPU cores (here 4) to use for each process is defined in the *.param file. With cores(args) one can return this value from the SYSargs object. Note, if a module system is not installed or used, then the corresponding *.param file needs to be edited accordingly by either providing an empty field in the line(s) starting with module or by deleting these lines.

bampaths <- runCommandline(args=args)

Alternatively, the computation can be greatly accelerated by processing many files in parallel using several compute nodes of a cluster, where a scheduling/queuing system is used for load balancing. To avoid over-subscription of CPU cores on the compute nodes, the value from cores(args) is passed on to the submission command, here nodes in the resources list object. The number of independent parallel cluster processes is defined under the Njobs argument. The following example will run 18 processes in parallel using for each 4 CPU cores. If the resources available on a cluster allow to run all 18 processes at the same time then the shown sample submission will utilize in total 72 CPU cores. Note, clusterRun can be used with most queueing systems as it is based on utilities from the BatchJobs package which supports the use of template files (*.tmpl) for defining the run parameters of different schedulers. To run the following code, one needs to have both a conf file (see .BatchJob samples here) and a template file (see *.tmpl samples here) for the queueing available on a system. The following example uses the sample conf and template files for the Torque scheduler provided by this package.

resources <- list(walltime="20:00:00", nodes=paste0("1:ppn=", cores(args)), memory="10gb")
reg <- clusterRun(args, conffile=".BatchJobs.R", template="torque.tmpl", Njobs=18, runid="01", 
                  resourceList=resources)
waitForJobs(reg)

Useful commands for monitoring progress of submitted jobs

showStatus(reg)
file.exists(outpaths(args))
sapply(1:length(args), function(x) loadResult(reg, x)) # Works after job completion
Back to Table of Contents

3.5 Read and alignment count stats

Generate table of read and alignment counts for all samples.

read_statsDF <- alignStats(args) 
write.table(read_statsDF, "results/alignStats.xls", row.names=FALSE, quote=FALSE, sep="\t")

The following shows the first four lines of the sample alignment stats file provided by the systemPipeR package. For simplicity the number of PE reads is multiplied here by 2 to approximate proper alignment frequencies where each read in a pair is counted.

read.table(system.file("extdata", "alignStats.xls", package="systemPipeR"), header=TRUE)[1:4,]
##   FileName Nreads2x Nalign Perc_Aligned Nalign_Primary Perc_Aligned_Primary
## 1      M1A   192918 177961     92.24697         177961             92.24697
## 2      M1B   197484 159378     80.70426         159378             80.70426
## 3      A1A   189870 176055     92.72397         176055             92.72397
## 4      A1B   188854 147768     78.24457         147768             78.24457

Parallelization of read/alignment stats on single machine with multiple cores

f <- function(x) alignStats(args[x])
read_statsList <- bplapply(seq(along=args), f, BPPARAM = MulticoreParam(workers=8))
read_statsDF <- do.call("rbind", read_statsList)

Parallelization of read/alignment stats via scheduler (e.g. Torque) across several compute nodes

library(BiocParallel); library(BatchJobs)
f <- function(x) {
    library(systemPipeR)
    args <- systemArgs(sysma="tophat.param", mytargets="targets.txt")
    alignStats(args[x])
}
funs <- makeClusterFunctionsTorque("torque.tmpl")
param <- BatchJobsParam(length(args), resources=list(walltime="20:00:00", nodes="1:ppn=1", memory="6gb"), cluster.functions=funs)
register(param)
read_statsList <- bplapply(seq(along=args), f)
read_statsDF <- do.call("rbind", read_statsList)
Back to Table of Contents

3.7 Alternative NGS Aligners

3.7.1 Alignment with Bowtie2 (e.g. for miRNA profiling)

The following example runs Bowtie2 as a single process without submitting it to a cluster.

args <- systemArgs(sysma="bowtieSE.param", mytargets="targets.txt")
moduleload(modules(args)) # Skip if module system is not available
bampaths <- runCommandline(args=args)

Alternatively, submit the job to compute nodes.

resources <- list(walltime="20:00:00", nodes=paste0("1:ppn=", cores(args)), memory="10gb")
reg <- clusterRun(args, conffile=".BatchJobs.R", template="torque.tmpl", Njobs=18, runid="01", 
                  resourceList=resources)
waitForJobs(reg)
Back to Table of Contents

3.7.2 Alignment with BWA-MEM (e.g. for VAR-Seq)

The following example runs BWA-MEM as a single process without submitting it to a cluster.

args <- systemArgs(sysma="param/bwa.param", mytargets="targets.txt")
moduleload(modules(args)) # Skip if module system is not available
system("bwa index -a bwtsw ./data/tair10.fasta") # Indexes reference genome
bampaths <- runCommandline(args=args[1:2])
Back to Table of Contents

3.7.3 Alignment with Rsubread (e.g. for RNA-Seq)

The following example shows how one can use within the environment the R-based aligner or other R-based functions that read from input files and write to output files.

library(Rsubread)
args <- systemArgs(sysma="param/rsubread.param", mytargets="targets.txt")
buildindex(basename=reference(args), reference=reference(args)) # Build indexed reference genome
align(index=reference(args), readfile1=infile1(args)[1:4], input_format="FASTQ", 
      output_file=outfile1(args)[1:4], output_format="SAM", nthreads=8, indels=1, TH1=2)
for(i in seq(along=outfile1(args))) asBam(file=outfile1(args)[i], destination=gsub(".sam", "", outfile1(args)[i]), overwrite=TRUE, indexDestination=TRUE)
Back to Table of Contents

3.7.4 Alignment with gsnap (e.g. for VAR-Seq and RNA-Seq)

Another R-based short read aligner is gsnap from the gmapR package (Wu and Nacu 2010). The code sample below introduces how to run this aligner on multiple nodes of a compute cluster.

library(gmapR); library(BiocParallel); library(BatchJobs)
args <- systemArgs(sysma="param/gsnap.param", mytargets="targetsPE.txt")
gmapGenome <- GmapGenome(reference(args), directory="data", name="gmap_tair10chr/", create=TRUE)
f <- function(x) {
    library(gmapR); library(systemPipeR)
    args <- systemArgs(sysma="gsnap.param", mytargets="targetsPE.txt")
    gmapGenome <- GmapGenome(reference(args), directory="data", name="gmap_tair10chr/", create=FALSE)
    p <- GsnapParam(genome=gmapGenome, unique_only=TRUE, molecule="DNA", max_mismatches=3)
    o <- gsnap(input_a=infile1(args)[x], input_b=infile2(args)[x], params=p, output=outfile1(args)[x])
}
funs <- makeClusterFunctionsTorque("torque.tmpl")
param <- BatchJobsParam(length(args), resources=list(walltime="20:00:00", nodes="1:ppn=1", memory="6gb"), cluster.functions=funs)
register(param)
d <- bplapply(seq(along=args), f)
Back to Table of Contents

3.8 Read counting for mRNA profiling experiments

Create txdb (needs to be done only once)

library(GenomicFeatures)
txdb <- makeTxDbFromGFF(file="data/tair10.gff", format="gff", dataSource="TAIR", organism="A. thaliana")
saveDb(txdb, file="./data/tair10.sqlite")
Back to Table of Contents

The following performs read counting with summarizeOverlaps in parallel mode with multiple cores.

library(BiocParallel)
txdb <- loadDb("./data/tair10.sqlite")
eByg <- exonsBy(txdb, by="gene")
bfl <- BamFileList(outpaths(args), yieldSize=50000, index=character())
multicoreParam <- MulticoreParam(workers=4); register(multicoreParam); registered()
counteByg <- bplapply(bfl, function(x) summarizeOverlaps(eByg, x, mode="Union", ignore.strand=TRUE, inter.feature=TRUE, singleEnd=TRUE)) # Note: for strand-specific RNA-Seq set 'ignore.strand=FALSE' and for PE data set 'singleEnd=FALSE'
countDFeByg <- sapply(seq(along=counteByg), function(x) assays(counteByg[[x]])$counts)
rownames(countDFeByg) <- names(rowRanges(counteByg[[1]])); colnames(countDFeByg) <- names(bfl)
rpkmDFeByg <- apply(countDFeByg, 2, function(x) returnRPKM(counts=x, ranges=eByg))
write.table(countDFeByg, "results/countDFeByg.xls", col.names=NA, quote=FALSE, sep="\t")
write.table(rpkmDFeByg, "results/rpkmDFeByg.xls", col.names=NA, quote=FALSE, sep="\t")

Please note, in addition to read counts this step generates RPKM normalized expression values. For most statistical differential expression or abundance analysis methods, such as edgeR or DESeq2, the raw count values should be used as input. The usage of RPKM values should be restricted to specialty applications required by some users, e.g. manually comparing the expression levels of different genes or features.

Back to Table of Contents

Read counting with summarizeOverlaps using multiple nodes of a cluster

library(BiocParallel)
f <- function(x) {
    library(systemPipeR); library(BiocParallel); library(GenomicFeatures)
    txdb <- loadDb("./data/tair10.sqlite")
    eByg <- exonsBy(txdb, by="gene")
    args <- systemArgs(sysma="tophat.param", mytargets="targets.txt")
    bfl <- BamFileList(outpaths(args), yieldSize=50000, index=character())
    summarizeOverlaps(eByg, bfl[x], mode="Union", ignore.strand=TRUE, inter.feature=TRUE, singleEnd=TRUE)
}
funs <- makeClusterFunctionsTorque("torque.tmpl")
param <- BatchJobsParam(length(args), resources=list(walltime="20:00:00", nodes="1:ppn=1", memory="6gb"), cluster.functions=funs)
register(param)
counteByg <- bplapply(seq(along=args), f) 
countDFeByg <- sapply(seq(along=counteByg), function(x) assays(counteByg[[x]])$counts)
rownames(countDFeByg) <- names(rowRanges(counteByg[[1]])); colnames(countDFeByg) <- names(outpaths(args))
Back to Table of Contents

3.9 Read counting for miRNA profiling experiments

Download miRNA genes from miRBase

system("wget ftp://mirbase.org/pub/mirbase/19/genomes/My_species.gff3 -P ./data/")
gff <- import.gff("./data/My_species.gff3")
gff <- split(gff, elementMetadata(gff)$ID)
bams <- names(bampaths); names(bams) <- targets$SampleName
bfl <- BamFileList(bams, yieldSize=50000, index=character())
countDFmiR <- summarizeOverlaps(gff, bfl, mode="Union", ignore.strand=FALSE, inter.feature=FALSE) # Note: inter.feature=FALSE important since pre and mature miRNA ranges overlap
rpkmDFmiR <- apply(countDFmiR, 2, function(x) returnRPKM(counts=x, gffsub=gff))
write.table(assays(countDFmiR)$counts, "results/countDFmiR.xls", col.names=NA, quote=FALSE, sep="\t")
write.table(rpkmDFmiR, "results/rpkmDFmiR.xls", col.names=NA, quote=FALSE, sep="\t")
Back to Table of Contents

3.10 Correlation analysis of samples

The following computes the sample-wise Spearman correlation coefficients from the rlog (regularized-logarithm) transformed expression values generated with the DESeq2 package. After transformation to a distance matrix, hierarchical clustering is performed with the hclust function and the result is plotted as a dendrogram (sample_tree.pdf).

library(DESeq2, warn.conflicts=FALSE, quietly=TRUE); library(ape, warn.conflicts=FALSE)
countDFpath <- system.file("extdata", "countDFeByg.xls", package="systemPipeR")
countDF <- as.matrix(read.table(countDFpath))
colData <- data.frame(row.names=targetsin(args)$SampleName, condition=targetsin(args)$Factor)
dds <- DESeqDataSetFromMatrix(countData = countDF, colData = colData, design = ~ condition)
d <- cor(assay(rlog(dds)), method="spearman")
hc <- hclust(dist(1-d))
plot.phylo(as.phylo(hc), type="p", edge.col=4, edge.width=3, show.node.label=TRUE, no.margin=TRUE)

Figure 3: Correlation dendrogram of samples for rlog values.


Alternatively, the clustering can be performed with RPKM normalized expression values. In combination with Spearman correlation the results of the two clustering methods are often relatively similar.

rpkmDFeBygpath <- system.file("extdata", "rpkmDFeByg.xls", package="systemPipeR")
rpkmDFeByg <- read.table(rpkmDFeBygpath, check.names=FALSE)
rpkmDFeByg <- rpkmDFeByg[rowMeans(rpkmDFeByg) > 50,]
d <- cor(rpkmDFeByg, method="spearman")
hc <- hclust(as.dist(1-d))
plot.phylo(as.phylo(hc), type="p", edge.col="blue", edge.width=2, show.node.label=TRUE, no.margin=TRUE)
Back to Table of Contents

3.11 DEG analysis with edgeR

The following run_edgeR function is a convenience wrapper for identifying differentially expressed genes (DEGs) in batch mode with edgeR’s GML method (Robinson, McCarthy, and Smyth 2010) for any number of pairwise sample comparisons specified under the cmp argument. Users are strongly encouraged to consult the edgeR vignette for more detailed information on this topic and how to properly run edgeR on data sets with more complex experimental designs.

targets <- read.delim(targetspath, comment="#")
cmp <- readComp(file=targetspath, format="matrix", delim="-")
cmp[[1]]
##       [,1]  [,2] 
##  [1,] "M1"  "A1" 
##  [2,] "M1"  "V1" 
##  [3,] "A1"  "V1" 
##  [4,] "M6"  "A6" 
##  [5,] "M6"  "V6" 
##  [6,] "A6"  "V6" 
##  [7,] "M12" "A12"
##  [8,] "M12" "V12"
##  [9,] "A12" "V12"
countDFeBygpath <- system.file("extdata", "countDFeByg.xls", package="systemPipeR")
countDFeByg <- read.delim(countDFeBygpath, row.names=1)
edgeDF <- run_edgeR(countDF=countDFeByg, targets=targets, cmp=cmp[[1]], independent=FALSE, mdsplot="")
## Disp = 0.20653 , BCV = 0.4545

Filter and plot DEG results for up and down regulated genes. Because of the small size of the toy data set used by this vignette, the FDR value has been set to a relatively high threshold (here 10%). More commonly used FDR cutoffs are 1% or 5%. The definition of ‘up’ and ‘down’ is given in the corresponding help file. To open it, type ?filterDEGs in the R console.

DEG_list <- filterDEGs(degDF=edgeDF, filter=c(Fold=2, FDR=10))
Figure 4: Up and down regulated DEGs identified by edgeR.


names(DEG_list)
## [1] "UporDown" "Up"       "Down"     "Summary"
DEG_list$Summary[1:4,]
##       Comparisons Counts_Up_or_Down Counts_Up Counts_Down
## M1-A1       M1-A1                 0         0           0
## M1-V1       M1-V1                 1         1           0
## A1-V1       A1-V1                 1         1           0
## M6-A6       M6-A6                 0         0           0
Back to Table of Contents

3.12 DEG analysis with DESeq2

The following run_DESeq2 function is a convenience wrapper for identifying DEGs in batch mode with DESeq2 (Love, Huber, and Anders 2014) for any number of pairwise sample comparisons specified under the cmp argument. Users are strongly encouraged to consult the DESeq2 vignette for more detailed information on this topic and how to properly run DESeq2 on data sets with more complex experimental designs.

degseqDF <- run_DESeq2(countDF=countDFeByg, targets=targets, cmp=cmp[[1]], independent=FALSE)

Filter and plot DEG results for up and down regulated genes.

DEG_list2 <- filterDEGs(degDF=degseqDF, filter=c(Fold=2, FDR=10))
Figure 5: Up and down regulated DEGs identified by DESeq2.


Back to Table of Contents

3.13 Venn Diagrams

The function overLapper can compute Venn intersects for large numbers of sample sets (up to 20 or more) and vennPlot can plot 2-5 way Venn diagrams. A useful feature is the possiblity to combine the counts from several Venn comparisons with the same number of sample sets in a single Venn diagram (here for 4 up and down DEG sets).

vennsetup <- overLapper(DEG_list$Up[6:9], type="vennsets")
vennsetdown <- overLapper(DEG_list$Down[6:9], type="vennsets")
vennPlot(list(vennsetup, vennsetdown), mymain="", mysub="", colmode=2, ccol=c("blue", "red"))
Figure 6: Venn Diagram for 4 Up and Down DEG Sets.


Back to Table of Contents

3.14 GO term enrichment analysis of DEGs

3.14.1 Obtain gene-to-GO mappings

The following shows how to obtain gene-to-GO mappings from biomaRt (here for A. thaliana) and how to organize them for the downstream GO term enrichment analysis. Alternatively, the gene-to-GO mappings can be obtained for many organisms from Bioconductor’s *.db genome annotation packages or GO annotation files provided by various genome databases. For each annotation this relatively slow preprocessing step needs to be performed only once. Subsequently, the preprocessed data can be loaded with the load function as shown in the next subsection.

library("biomaRt")
listMarts() # To choose BioMart database
m <- useMart("ENSEMBL_MART_PLANT"); listDatasets(m) 
m <- useMart("ENSEMBL_MART_PLANT", dataset="athaliana_eg_gene")
listAttributes(m) # Choose data types you want to download
go <- getBM(attributes=c("go_accession", "tair_locus", "go_namespace_1003"), mart=m)
go <- go[go[,3]!="",]; go[,3] <- as.character(go[,3])
dir.create("./data/GO")
write.table(go, "data/GO/GOannotationsBiomart_mod.txt", quote=FALSE, row.names=FALSE, col.names=FALSE, sep="\t")
catdb <- makeCATdb(myfile="data/GO/GOannotationsBiomart_mod.txt", lib=NULL, org="", colno=c(1,2,3), idconv=NULL)
save(catdb, file="data/GO/catdb.RData") 
Back to Table of Contents

3.14.2 Batch GO term enrichment analysis

Apply the enrichment analysis to the DEG sets obtained in the above differential expression analysis. Note, in the following example the FDR filter is set here to an unreasonably high value, simply because of the small size of the toy data set used in this vignette. Batch enrichment analysis of many gene sets is performed with the GOCluster_Report function. When method="all", it returns all GO terms passing the p-value cutoff specified under the cutoff arguments. When method="slim", it returns only the GO terms specified under the myslimv argument. The given example shows how one can obtain such a GO slim vector from BioMart for a specific organism.

load("data/GO/catdb.RData")
DEG_list <- filterDEGs(degDF=edgeDF, filter=c(Fold=2, FDR=50), plot=FALSE)
up_down <- DEG_list$UporDown; names(up_down) <- paste(names(up_down), "_up_down", sep="")
up <- DEG_list$Up; names(up) <- paste(names(up), "_up", sep="")
down <- DEG_list$Down; names(down) <- paste(names(down), "_down", sep="")
DEGlist <- c(up_down, up, down)
DEGlist <- DEGlist[sapply(DEGlist, length) > 0]
BatchResult <- GOCluster_Report(catdb=catdb, setlist=DEGlist, method="all", id_type="gene", CLSZ=2, cutoff=0.9, gocats=c("MF", "BP", "CC"), recordSpecGO=NULL)
library("biomaRt"); m <- useMart("ENSEMBL_MART_PLANT", dataset="athaliana_eg_gene")
goslimvec <- as.character(getBM(attributes=c("goslim_goa_accession"), mart=m)[,1])
BatchResultslim <- GOCluster_Report(catdb=catdb, setlist=DEGlist, method="slim", id_type="gene", myslimv=goslimvec, CLSZ=10, cutoff=0.01, gocats=c("MF", "BP", "CC"), recordSpecGO=NULL)
Back to Table of Contents

3.14.3 Plot batch GO term results

The data.frame generated by GOCluster_Report can be plotted with the goBarplot function. Because of the variable size of the sample sets, it may not always be desirable to show the results from different DEG sets in the same bar plot. Plotting single sample sets is achieved by subsetting the input data frame as shown in the first line of the following example.

gos <- BatchResultslim[grep("M6-V6_up_down", BatchResultslim$CLID), ]
gos <- BatchResultslim
pdf("GOslimbarplotMF.pdf", height=8, width=10); goBarplot(gos, gocat="MF"); dev.off()
goBarplot(gos, gocat="BP")
goBarplot(gos, gocat="CC")
Figure 7: GO Slim Barplot for MF Ontology.


Back to Table of Contents

3.15 Clustering and heat maps

The following example performs hierarchical clustering on the rlog transformed expression matrix subsetted by the DEGs identified in the above differential expression analysis. It uses a Pearson correlation-based distance measure and complete linkage for cluster joining.

library(pheatmap)
geneids <- unique(as.character(unlist(DEG_list[[1]])))
y <- assay(rlog(dds))[geneids, ]
pdf("heatmap1.pdf")
pheatmap(y, scale="row", clustering_distance_rows="correlation", clustering_distance_cols="correlation")
dev.off()
Figure 8: Heat map with hierarchical clustering