%\VignetteIndexEntry{GenomicRanges HOWTOs}
%\VignetteDepends{GenomicRanges, Rsamtools, pasillaBamSubset, TxDb.Dmelanogaster.UCSC.dm3.ensGene, AnnotationHub, DESeq, edgeR, TxDb.Hsapiens.UCSC.hg19.knownGene, GenomicFeatures, Biostrings, BSgenome.Hsapiens.UCSC.hg19, KEGG.db, KEGGgraph, BSgenome.Scerevisiae.UCSC.sacCer2}
%\VignetteKeywords{sequence, sequencing, alignments}
%\VignettePackage{GenomicRanges}

\documentclass{article}

\usepackage[authoryear,round]{natbib}
\bibliographystyle{plainnat}

<<style, eval=TRUE, echo=FALSE, results=tex>>=
BiocStyle::latex()
@

\title{\Biocpkg{GenomicRanges} HOWTOs}
\author{Bioconductor Team}
\date{Edited: October 2013; Compiled: \today}

\begin{document}

\maketitle

\tableofcontents

<<options, echo=FALSE>>=
options(width=72)
options("showHeadLines" = 3)
options("showTailLines" = 3)
@


%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

\section{Introduction}

This vignette is a collection of {\it HOWTOs}. Each {\it HOWTO} is
a short section that demonstrates how to use the containers and
operations implemented in the \Biocpkg{GenomicRanges} and related
packages (\Biocpkg{IRanges}, \Biocpkg{GenomicFeatures},
\Biocpkg{Rsamtools}, and \Biocpkg{Biostrings}) to perform a task
typically found in the context of a high throughput sequence analysis.

The HOWTOs are self contained, independent of each other, and can be
studied and reproduced in any order.

We assume the reader has some previous experience with \R{} and
with basic manipulation of \Rcode{GRanges}, \Rcode{GRangesList}, \Rcode{Rle},
\Rcode{RleList}, and \Rcode{DataFrame} objects. See the ``An Introduction
to Genomic Ranges Classes'' vignette located in the \Biocpkg{GenomicRanges}
package (in the same folder as this document) for an introduction to these
containers.

Additional recommended readings after this vignette are the ``Software for
Computing and Annotating Genomic Ranges'' paper[\citet{Lawrence2013ranges}]
and the ``Counting reads with \Rfunction{summarizeOverlaps}'' vignette
located in the \Biocpkg{GenomicRanges} package (in the same folder as this
document).

To display the list of vignettes available in the \Biocpkg{GenomicRanges},
use \Rcode{browseVignettes("GenomicRanges")}.


%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

\section{How to read BAM files into \R{}}

As sample data we use the \Biocpkg{pasillaBamSubset} data package which
contains both a BAM file with single-end reads (untreated1\_chr4) and
a BAM file with paired-end reads (untreated3\_chr4). Each file is a subset
of chr4 from the "Pasilla" experiment. See ?\Biocpkg{pasillaBamSubset}
for details.

<<load, results=hide>>=
library(GenomicRanges)
library(Rsamtools)
library(pasillaBamSubset)
un1 <- untreated1_chr4() ## single-end reads
@

Several functions are available for reading BAM files into \R{}:

\begin{verbatim}
  scanBam()
  readGAlignments()
  readGAlignmentPairs()
  readGAlignmentsList()
\end{verbatim}

\Rfunction{scanBam} is a low-level function that returns a list of lists
and is not discussed further here. For details see ?\Rfunction{scanBam}.

\subsection{Single-end reads}
Single-end reads can be loaded with the \Rfunction{readGAlignments} function.

<<readGAlignments>>=
un1 <- untreated1_chr4()
gal <- readGAlignments(un1)
@

Data subsets can be specified by genomic position, field names, or flag
criteria in the \Rcode{ScanBamParam}. Here we input records that overlap 
position 1 to 5000 on the negative strand with \Rcode{flag} and 
\Rcode{cigar} as metadata columns.
 
<<ScanBamParam>>=
what <- c("flag", "cigar") 
which <- GRanges("chr4", IRanges(1, 5000)) 
flag <- scanBamFlag(isMinusStrand = TRUE)
param <- ScanBamParam(which=which, what=what, flag=flag)
neg <- readGAlignments(un1, param=param)
neg
@

Another approach to subsetting the data is to use \Rfunction{filterBam}.
This function creates a new BAM file of records passing user-defined 
criteria. See ?\Rfunction{filterBam} for details.

\subsection{Paired-end reads}

Paired-end reads can be loaded with \Rfunction{readGAlignmentPairs}
or \Rfunction{readGAlignmentsList}. These functions use the same
mate paring algorithm but output different objects.

Let's start with \Rfunction{readGAlignmentPairs}:

<<readGAlignmentPairs>>=
un3 <- untreated3_chr4()
gapairs <- readGAlignmentPairs(un3)
@

The \Robject{GAlignmentPairs} class holds only pairs; reads with no
mate or with ambiguous pairing are discarded.
Each list element holds exactly 2 records (a mated pair). Records
can be accessed as the \Rcode{first} and\Rcode{last} segments in
a template or as \Rcode{left} and \Rcode{right} alignments.
See ?\Rfunction{GAlignmentPairs} for details.

<<gapairs>>=
gapairs
@ 

For \Rcode{readGAlignmentsList}, mate pairing is performed when \Rcode{asMates}
is set to \Rcode{TRUE} on the \Rcode{BamFile} object, otherwise records are
treated as single-end. 

<<readGAlignmentsList>>=
galist <- readGAlignmentsList(BamFile(un3, asMates=TRUE))
@

\Robject{GAlignmentsList} is a more general `list-like' structure
that holds mate pairs as well as non-mates (i.e., singletons, records 
with unmapped mates etc.) A \Rcode{mates} metadata column (accessed
with \Rfunction{mcols}) indicates which records were paired and is set 
on both the individual \Robject{GAlignments} and the outer list elements.

<<galist>>=
galist
@

Non-mated reads are returned as groups by QNAME and contain any number 
of records. Here the non-mate groups range in size from 1 to 9.

<<non_mates>>=
non_mates <- galist[unlist(mcols(galist)$mates) == FALSE]
table(elementLengths(non_mates))
@

\subsection{Iterating with \Rcode{yieldSize}}

Large files can be iterated through in chunks by setting a \Rcode{yieldSize} 
on the \Rcode{BamFile}.

<<yieldSize>>=
bf <- BamFile(un1, yieldSize=100000)
@

Iteration through a BAM file requires that the file be opened, repeatedly
queried inside a loop, then closed. Repeated calls to 
\Rfunction{readGAlignments} without opening the file first result
in the same 100000 records returned each time.

<<readGAlignments_by_chunk>>=
open(bf)
cvg <- NULL
repeat {
    chunk <- readGAlignments(bf)
    if (length(chunk) == 0L)
        break
    chunk_cvg <- coverage(chunk)
    if (is.null(cvg)) {
        cvg <- chunk_cvg
    } else {
        cvg <- cvg + chunk_cvg
    }
}
close(bf)
cvg
@


%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

\section{How to prepare a table of read counts for RNA-Seq differential
         gene expression}

Methods for RNA-Seq gene expression analysis generally require a table of
counts that summarize the number of reads that overlap or `hit' a 
particular gene. In this section we count with \Rcode{summarizeOverlaps}
and create a count table from the results. 

Other packages that provide read counting are \Biocpkg{Rsubread} and 
\Biocpkg{easyRNASeq}. The \Biocpkg{parathyroidSE} package vignette 
contains a workflow on counting and other common operations required for 
differential expression analysis. 

\subsection{Counting with \Rfunction{summarizeOverlaps}}

As sample data we use \Biocpkg{pasillaBamSubset} which contains both
a single-end BAM (untreated1\_chr4) and a paired-end BAM
(untreated3\_chr4). Each file is a subset of chr4 from the "Pasilla" 
experiment. See ?\Biocpkg{pasillaBamSubset} for details.

<<load, results=hide>>=
library(GenomicRanges)
library(Rsamtools)
library(pasillaBamSubset)
un1 <- untreated1_chr4() ## single-end records
@

\Rcode{summarizeOverlaps} requires the name of a BAM file(s) and an
annotation to count against. The annotation must match the genome 
build the BAM records were aligned to. For the pasilla data this is 
dm3 Dmelanogaster which is available as a \Bioconductor{} package. 
Load the package and extract the exon ranges by gene.

<<count_1>>=
library(TxDb.Dmelanogaster.UCSC.dm3.ensGene)
exbygene <- exonsBy(TxDb.Dmelanogaster.UCSC.dm3.ensGene, "gene")
@

\Rcode{summarizeOverlaps} automatically sets a \Rcode{yieldSize}
on large BAM files and iterates over them in chunks. When reading
paired-end data set the \Rcode{singleEnd} argument to FALSE.
See ?\Rfunction{summarizeOverlaps} for details reguarding the
count \Rcode{modes} and additional arguments. 

<<count_2>>=
se <- summarizeOverlaps(exbygene, un1, mode="IntersectionNotEmpty")
@

The return object is a \Rcode{SummarizedExperiment} with counts in 
the \Rcode{assays} slot.

<<count_3>>=
class(se)
head(table(assays(se)$counts))
@

The count vector is the same length as the annotation.

<<count_4>>=
identical(length(exbygene), length(assays(se)$counts))
@

The annotation is stored in the \Rcode{rowData} slot.

<<count_5>>=
rowData(se)
@

\subsection{Retrieving annotations from \Biocpkg{AnnotationHub}}

When the annotation is not available as a \Rcode{GRanges} or a
\Bioconductor{} package it may be available in \Rcode{AnnotationHub}.
Create a `hub' and filter on Drosophila melanogaster.

<<hub_1>>=
library(AnnotationHub)
hub <- AnnotationHub()
filters(hub) <- list(Species="Drosophila melanogaster")
@

There are 87 files that match Drosophila melanogaster.

<<hub_2>>=
length(hub)
head(names(hub))
@

Retrieve a dm3 file as a \Rcode{GRanges}. 

<<hub_3>>=
gr <- hub$goldenpath.dm3.database.ensGene_0.0.1.RData
summary(gr)
@

The metadata fields contain the details of file origin and content.

<<hub_4>>=
names(metadata(gr)[[2]])
metadata(gr)[[2]]$Tags
@

Split the GRanges by gene name to get a \Robject{GRangesList} of
transcripts by gene.

<<hub_5>>= 
split(gr, gr$name)
@

Before performing overlap operations confirm that the seqlevels 
(chromosome names) in the annotation match those in the BAM file. See 
?\Rfunction{renameSeqlevels}, ?\Rfunction{keepSeqlevels} and 
?\Rfunction{seqlevels} for examples of renaming seqlevels.

\subsection{Count tables}

Two popular packages for gene expression are \Biocpkg{DESeq} and 
\Biocpkg{edgeR}. Tables of counts per gene are required for both and can be 
easily created with a vector of counts. Here we use the counts from the 
\Robject{SummarizedExperiment}.

<<count_table>>=
library(DESeq)
deseq <- newCountDataSet(assays(se)$counts, rownames(colData(se)))
library(edgeR)
edger <- DGEList(assays(se)$counts, group=rownames(colData(se)))
@

 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

\section{How to extract DNA sequences of gene regions}

\subsection{DNA sequences for intron and exon regions of a single gene}

DNA sequences for the introns and exons of a gene are essentially
the sequences for the introns and exons for all known transcripts of
a gene. The first task is to identify all transcripts associated
with the gene of interest. Our sample gene is the human TRAK2
which is involved in regulation of endosome-to-lysosome trafficking 
of membrane cargo. The Entrez gene id is `66008'. 

<<trak_1>>=
trak2 <- "66008"
@

Load the UCSC `Known Gene' table annotation available as a
\Bioconductor{} package.

<<trak_2>>=
library(TxDb.Hsapiens.UCSC.hg19.knownGene)
txdb <- TxDb.Hsapiens.UCSC.hg19.knownGene
@

To get the transcripts associated with the trak2 gene we use the
\Rfunction{transcriptsBy} function from the \Biocpkg{GenomicFeatures} 
package. This returns a \Robject{GRangesList} of all transcripts
grouped by gene. We are only interested in trak2 so we subset the 
list on the trak2 gene id.

<<trak_3>>=
library(GenomicFeatures)
txbygene <- transcriptsBy(txdb, by="gene")[trak2]
txbygene
@

The transcript names corresponding to the trak2 gene will be used to 
subset the extracted intron and exon regions. The \Rcode{txbygene} 
object is a \Robject{GRangesList} and the transcript names are a 
metadata column on the individual \Robject{GRanges}. To extract the 
names we must first `flatten' or unlist \Rcode{txbygene}.

<<trak_4>>=
tx_names <- mcols(unlist(txbygene))$tx_name
tx_names
@

Intron and exon regions are extracted with \Rfunction{intronsByTranscript} 
and \Rfunction{exonsBy}. The resulting \Robject{GRangesLists} are
subset on the trak2 transcript names. 

Extract the intron regions ...

<<trak_5>>=
intronsbytx <- intronsByTranscript(txdb, use.names=TRUE)[tx_names]
elementLengths(intronsbytx)
@

and the exon regions.

<<trak_7>>=
exonsbytx <- exonsBy(txdb, "tx", use.names=TRUE)[tx_names]
elementLengths(exonsbytx)
@

Next we want the DNA sequences for these intron and exon regions.
The \Rfunction{extractTranscriptsFromGenome} function in the 
\Biocpkg{Biostrings} package will query a \Biocpkg{BSGenome} package 
with a set of genomic positions and retrieve the DNA sequences. 

<<trak_8>>=
library(Biostrings)
library(BSgenome.Hsapiens.UCSC.hg19)
@

Extract the intron sequences ...

<<trak_9>>=
intron_seqs <- extractTranscriptsFromGenome(Hsapiens, intronsbytx)
intron_seqs
@

and the exon sequences.

<<trak_10>>=
exon_seqs <- extractTranscriptsFromGenome(Hsapiens, exonsbytx)
exon_seqs
@


\subsection{DNA sequences for coding and UTR regions of genes associated 
            with colorectal cancer}

In this section we extract the coding and UTR sequences of genes involved
in colorectal cancer. The workflow extends the ideas presented in the single 
gene example and suggests an approach to identify disease-related genes.

\subsubsection{Build a gene list}

We start with a list of gene or transcript ids. If you do not have 
pre-defined list one can be created with the \Biocpkg{KEGG.db} and 
\Biocpkg{KEGGgraph} packages. Updates to the data in the \Biocpkg{KEGG.db} 
package are no longer available, however, the resource is still useful for 
identifying pathway names and ids. 

Create a table of KEGG pathways and ids and search on the term `cancer'.

<<cancer_1>>=
library(KEGG.db)
pathways <- toTable(KEGGPATHNAME2ID)
pathways[grepl("cancer", pathways$path_name, fixed=TRUE),] 
@

Use the "05210" id to query the KEGG web resource (accesses the currently
maintained data).

<<cancer_2>>=
library(KEGGgraph)
dest <- tempfile()
retrieveKGML("05200", "hsa", dest, "internal")
@

The suffix of the KEGG id is the Entrez gene id. The 
\Rfunction{translateKEGGID2GeneID} simply removes the prefix leaving 
just the Entrez gene ids.

<<cancer_3>>=
crids <- as.character(parseKGML2DataFrame(dest)[,1])
crgenes <- unique(translateKEGGID2GeneID(crids))
head(crgenes)
@

\subsubsection{Identify genomic coordinates}

The list of gene ids is used to extract genomic positions of the regions 
of interest. The Known Gene table from UCSC will be the annotation and
is available as a \Bioconductor{} package.

<<cancer_4>>=
library(TxDb.Hsapiens.UCSC.hg19.knownGene)
txdb <- TxDb.Hsapiens.UCSC.hg19.knownGene
@

If an annotation is not available as a \Bioconductor{} annotation package
it may be available in \Biocpkg{AnnotationHub}. Additionally, there are
functions in \Biocpkg{GenomicFeatures} which can retrieve data from UCSC and
Ensembl to create a \Robject{TranscriptDb}. See 
?\Rcode{makeTranscriptDbFromUCSC} for details.

As in the single gene example we need to identify the transcripts 
corresponding to each gene. The transcript id (or name) is used
to isolate the UTR and coding regions of interest. This grouping of
transcript by gene is also used to re-group the final sequence results.

The \Rcode{transcriptsBy} function outputs both the gene and transcript
identifiers which we use to create a map between the two. The 
\Rcode{map} is a \Robject{CharacterList} with gene ids as names and 
transcript ids as the list elements.

<<cancer_5>>=
txbygene <- transcriptsBy(txdb, "gene")[crgenes] ## subset on colorectal genes
map <- relist(unlist(txbygene, use.names=FALSE)$tx_id, txbygene)
map
@

Extract the UTR and coding regions.

<<cancer_6>>=
cds <- cdsBy(txdb, "tx")
threeUTR <- threeUTRsByTranscript(txdb)
fiveUTR <- fiveUTRsByTranscript(txdb)
@

Coding and UTR regions may not be present for all transcripts specified 
in \Rcode{map}. Consequently, the subset results will not be the same 
length. This length discrepancy must be taken into account when re-listing 
the final results by gene.

<<cancer_7>>=
txid <- unlist(map, use.names=FALSE)
cds <- cds[names(cds) %in% txid]
threeUTR <- threeUTR[names(threeUTR) %in% txid]
fiveUTR <- fiveUTR[names(fiveUTR) %in% txid]
@

Note the different lengths of the subset regions.

<<cancer_8>>=
length(txid) ## all possible transcripts
length(cds)
length(threeUTR)
length(fiveUTR)
@

These objects are \Robject{GRangesList}s with the transcript id as the 
outer list element. 

<<cancer_9>>=
cds
@

\subsubsection{Extract sequences from BSgenome}

The \Rcode{BSgenome} packages contain complete genome sequences
for a given organism.

Load the \Rcode{BSgenome} package for homo sapiens.

<<cancer_10>>=
library(BSgenome.Hsapiens.UCSC.hg19)
genome <- BSgenome.Hsapiens.UCSC.hg19
@

Use \Rfunction{extractTranscriptsFromGenome} to extract the UTR and coding 
regions from the \Rcode{BSGenome}. This function retrieves the sequences 
for an any \Robject{GRanges} or \Robject{GRangesList} (i.e., not just 
transcripts like the name implies).

<<cancer_11>>=
threeUTR_seqs <- extractTranscriptsFromGenome(genome, threeUTR) 
fiveUTR_seqs <- extractTranscriptsFromGenome(genome, fiveUTR) 
cds_seqs <- extractTranscriptsFromGenome(genome, cds) 
@

The return values are \Robject{DNAStringSet} objects.

<<cancer_12>>=
cds_seqs
@

Our final step is to collect the coding and UTR regions (currently 
organzied by transcript) into groups by gene id. The \Rfunction{split} 
function splits the sequences in the \Robject{DNAStringSet} by the partition
object. The partition object represents the number of transcript ranges 
(defined as the width) in each gene id group. These widths are different for 
each region because not all transcripts had a coding or 3' or 5' UTR region 
defined. 

<<cancer_13>>=
lst3 <- split(threeUTR_seqs, PartitioningByWidth(sum(map %in% names(threeUTR))))
lst5 <- split(fiveUTR_seqs, PartitioningByWidth(sum(map %in% names(fiveUTR))))
lstc <- split(cds_seqs, PartitioningByWidth(sum(map %in% names(cds))))
names(lst3) <- names(lst5) <- names(lstc) <- names(map)
@

There are 239 genes in \Rcode{map} each of which have 1 or more transcripts. 
The table of element lengths shows how many genes have each number of
transcripts. For example, 47 genes have 1 transcript, 48 genes have 2 etc.

<<cancer_14>>=
length(map)
table(elementLengths(map))
@

The lists of DNA sequences all have the same length as \Rcode{map} but one or
more of the element lengths may be zero. This would indicate that data were
not available for that gene. The tables below show that there was at least
1 coding region available for all genes (i.e., none of the element lengths
are 0). However, both the 3' and 5' UTR results have element lengths of 0 
which indicates no UTR data were available for that gene.

<<cancer_15>>=
table(elementLengths(lstc))
table(elementLengths(lst3))
names(lst3)[elementLengths(lst3) == 0L] ## genes with no 3' UTR data
table(elementLengths(lst5))
names(lst5)[elementLengths(lst5) == 0L] ## genes with no 5' UTR data
@


%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

\section{How to create DNA consensus sequences for read group `families'}

The motivation for this HOWTO comes from a study which explored the
dynamics of point mutations. The mutations of interest exist with a range 
of frequencies in the control group (e.g., 0.1\% - 50\%). PCR and sequencing 
error rates make it difficult to identify low frequency events 
(e.g., < 20\%).

When a library is prepared with Nextera, random fragments are generated 
followed by a few rounds of PCR. When the genome is large enough, reads 
aligning to the same start position are likely descendant from the same 
template fragment and should have identical sequences. 

The goal is to elimininate noise by grouping the reads by common start 
position and discarding those that do not exceed a certain threshold within 
each family. A new consensus sequence will be created for each read group
family.

\subsection{Sort reads into groups by start position}

Load the BAM file into a GAlignments object.

<<cseq_1>>=
library(Rsamtools)
bamfile <- system.file("extdata", "ex1.bam", package="Rsamtools")
param <- ScanBamParam(what=c("seq", "qual"))
gal <- readGAlignmentsFromBam(bamfile, use.names=TRUE, param=param)
@

Use the \Rfunction{sequenceLayer} function to {\it lay} the query sequences
and quality strings on the reference.

<<cseq_2>>=
qseq <- setNames(mcols(gal)$seq, names(gal))
qual <- setNames(mcols(gal)$qual, names(gal))
qseq_on_ref <- sequenceLayer(qseq, cigar(gal),
                             from="query", to="reference")
qual_on_ref <- sequenceLayer(qual, cigar(gal),
                             from="query", to="reference")
@

Split by chromosome.

<<cseq_3>>=
qseq_on_ref_by_chrom <- splitAsList(qseq_on_ref, seqnames(gal))
qual_on_ref_by_chrom <- splitAsList(qual_on_ref, seqnames(gal))
pos_by_chrom <- splitAsList(start(gal), seqnames(gal))
@

For each chromosome generate one GRanges object that contains
unique alignment start positions and attach 3 metadata columns
to it: the number of reads, the query sequences, and the quality
strings.

<<cseq_4>>=
gr_by_chrom <- lapply(seqlevels(gal),
  function(seqname)
  {
    qseq_on_ref2 <- qseq_on_ref_by_chrom[[seqname]]
    qual_on_ref2 <- qual_on_ref_by_chrom[[seqname]]
    pos2 <- pos_by_chrom[[seqname]]
    qseq_on_ref_per_pos <- split(qseq_on_ref2, pos2)
    qual_on_ref_per_pos <- split(qual_on_ref2, pos2)
    nread <- elementLengths(qseq_on_ref_per_pos)
    gr_mcols <- DataFrame(nread=unname(nread),
                          qseq_on_ref=unname(qseq_on_ref_per_pos),
                          qual_on_ref=unname(qual_on_ref_per_pos))
    gr <- GRanges(Rle(seqname, nrow(gr_mcols)),
                  IRanges(as.integer(names(nread)), width=1))
    mcols(gr) <- gr_mcols
    seqlevels(gr) <- seqlevels(gal)
    gr
  })
@

Combine all the GRanges objects obtained in (4) in 1 big GRanges
object:

<<cseq_5>>=
gr <- do.call(c, gr_by_chrom)
seqinfo(gr) <- seqinfo(gal)
@

`gr' is a GRanges object that contains unique alignment start positions:

<<cseq_6>>=
gr[1:6]
@

Look at qseq\_on\_ref and qual\_on\_ref.

<<cseq_7>>= 
qseq_on_ref
qual_on_ref
@

2 reads align to start position 13. Let's have a close look at their 
sequences:

<<cseq_8>>=
mcols(gr)$qseq_on_ref[[6]]
@

and their qualities:

<<cseq_9>>=
mcols(gr)$qual_on_ref[[6]]
@

Note that the sequence and quality strings are those projected to the 
reference so the first letter in those strings are on top of start 
position 13, the 2nd letter on top of position 14, etc...

\subsection{Remove low frequency reads}

For each start position, remove reads with and under-represented sequence 
(e.g. threshold = 20\% for the data used here which is low coverage).
A unique number is assigned to each unique sequence. This will make
future calculations easier and a little bit faster.

<<cseq_10>>=
qseq_on_ref <- mcols(gr)$qseq_on_ref
tmp <- unlist(qseq_on_ref, use.names=FALSE)
qseq_on_ref_id <- relist(match(tmp, tmp), qseq_on_ref)
@

Quick look at `qseq\_on\_ref\_id':
It's an IntegerList object with the same length and "shape"
as `qseq\_on\_ref'.

<<cseq_11>>=
qseq_on_ref_id
@

Remove the under represented ids from each list element of `qseq\_on\_ref\_id':

<<cseq_12>>=
qseq_on_ref_id2 <- endoapply(qseq_on_ref_id,
    function(ids) ids[countMatches(ids, ids) >= 0.2 * length(ids)])
@

Remove corresponding sequences from `qseq\_on\_ref':

<<cseq_13>>=
tmp <- unlist(qseq_on_ref_id2, use.names=FALSE)
qseq_on_ref2 <- relist(unlist(qseq_on_ref, use.names=FALSE)[tmp],
                       qseq_on_ref_id2)
@

\subsection{Create a consensus sequence for each read group family}

Compute 1 consensus matrix per chromosome:

<<cseq_14>>=
split_factor <- rep.int(seqnames(gr), elementLengths(qseq_on_ref2))
qseq_on_ref2 <- unlist(qseq_on_ref2, use.names=FALSE)
qseq_on_ref2_by_chrom <- splitAsList(qseq_on_ref2, split_factor)
qseq_pos_by_chrom <- splitAsList(start(gr), split_factor)

cm_by_chrom <- lapply(names(qseq_pos_by_chrom),
    function(seqname)
        consensusMatrix(qseq_on_ref2_by_chrom[[seqname]],
                        as.prob=TRUE,
                        shift=qseq_pos_by_chrom[[seqname]]-1,
                        width=seqlengths(gr)[[seqname]]))
names(cm_by_chrom) <- names(qseq_pos_by_chrom)
@

'cm\_by\_chrom' is a list of consensus matrices. Each matrix has 17 rows 
(1 per letter in the DNA alphabet) and 1 column per chromosome position.

<<cseq_15>>=
lapply(cm_by_chrom, dim)
@

Compute the consensus string from each consensus matrix. We'll put "+" 
in the strings wherever there is no coverage for that position, and "N" 
where there is coverage but no consensus.

<<cseq_16>>=
cs_by_chrom <- lapply(cm_by_chrom,
    function(cm) {
        ## need to "fix" 'cm' because consensusString()
        ## doesn't like consensus matrices with columns
        ## that contain only zeroes (e.g., chromosome
        ## positions with no coverage)
        idx <- colSums(cm) == 0L
        cm["+", idx] <- 1
        DNAString(consensusString(cm, ambiguityMap="N"))
    })
@

The new consensus strings.

<<cseq_17>>=
cs_by_chrom
@


%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

\section{How to compute binned averages along a genome}

In some applications, there is the need to compute the average of a 
variable along a genome for a set of predefined fixed-width regions 
(sometimes called "bins"). One such example is coverage.
Coverage is an \Robject{RleList} with one list element per
chromosome. Here we simulate a coverage list.

<<bim_1>>=
library(BSgenome.Scerevisiae.UCSC.sacCer2)
set.seed(22)
cov <- RleList(
    lapply(seqlengths(Scerevisiae),
           function(len) Rle(sample(-10:10, len, replace=TRUE))),
    compress=FALSE)
head(cov, 3)
@

Use the \Rfunction{tileGenome} function to create a set of bins along
the genome.

<<bin_2>>=
bins1 <- tileGenome(seqinfo(Scerevisiae), tilewidth=100,
                    cut.last.tile.in.chrom=TRUE)
@

We define the following function to compute the binned average of a
numerical variable defined along a genome.

\begin{verbatim}
Arguments:
  'bins': a GRanges object representing the genomic bins.
       Typically obtained by calling tileGenome() with
       'cut.last.tile.in.chrom=TRUE'.
  'numvar': a named RleList object representing a numerical
       variable defined along the genome covered by 'bins', which
       is the genome described by 'seqinfo(bins)'.
  'mcolname': the name to give to the metadata column that will
       contain the binned average in the returned object.
\end{verbatim}

The function returns `bins' with an additional metadata column named 
`mcolname' containing the binned average.

<<bin_3>>=
binnedAverage <- function(bins, numvar, mcolname)
{
    stopifnot(is(bins, "GRanges"))
    stopifnot(is(numvar, "RleList"))
    stopifnot(identical(seqlevels(bins), names(numvar)))
    bins_per_chrom <- split(ranges(bins), seqnames(bins))
    means_list <- lapply(names(numvar),
        function(seqname) {
            views <- Views(numvar[[seqname]],
                           bins_per_chrom[[seqname]])
            viewMeans(views)
        })
    new_mcol <- unsplit(means_list, as.factor(seqnames(bins)))
    mcols(bins)[[mcolname]] <- new_mcol
    bins
}
@

Compute the binned average for `cov':

<<bin_4>>=
bins1 <- binnedAverage(bins1, cov, "binned_cov")
bins1
@

The bin size can be modified with the \Rcode{tilewidth} argument
to \Rfunction{tileGenome}. For additional examples see
?\Rfunction{tileGenome}.

\section{Session Information}

<<SessionInfo, echo=FALSE>>=
sessionInfo()
@

\bibliography{GenomicRanges}

\end{document}