mapToTranscripts {GenomicFeatures} | R Documentation |
Map range coordinates between features in the transcriptome and genome (reference) space.
See ?mapToAlignments
in the
GenomicAlignments package for mapping coordinates between
reads (local) and genome (reference) space using a CIGAR alignment.
## mapping to transcripts ## S4 method for signature 'GenomicRanges,GenomicRanges' mapToTranscripts(x, transcripts, ignore.strand = FALSE) ## S4 method for signature 'GenomicRanges,GRangesList' mapToTranscripts(x, transcripts, ignore.strand = FALSE, intronJunctions=FALSE) ## S4 method for signature 'ANY,TxDb' mapToTranscripts(x, transcripts, ignore.strand = FALSE, extractor.fun = GenomicFeatures::transcripts, ...) ## S4 method for signature 'GenomicRanges,GRangesList' pmapToTranscripts(x, transcripts, ignore.strand = FALSE) ## mapping from transcripts ## S4 method for signature 'GenomicRanges,GRangesList' mapFromTranscripts(x, transcripts, ignore.strand = FALSE) ## S4 method for signature 'GenomicRanges,GRangesList' pmapFromTranscripts(x, transcripts, ignore.strand = FALSE) ## S4 method for signature 'IntegerRanges,GRangesList' pmapFromTranscripts(x, transcripts)
x |
GenomicRanges object of positions to be mapped.
The seqnames of |
transcripts |
A named GenomicRanges or
GRangesList object used to map between The |
ignore.strand |
When When Mapped position is computed by counting from the transcription start site
(TSS) and is not affected by the value of |
intronJunctions |
Logical to indicate whether intronic ranges in Ranges in This argument is only supported in |
extractor.fun |
Function to extract genomic features from a This argument is only applicable to Valid
|
... |
Additional arguments passed to |
In GenomicFeatures >= 1.21.10, the default for ignore.strand
was
changed to FALSE
for consistency with other methods in the
GenomicRanges and GenomicAlignments packages. Additionally,
the mapped position is computed from the TSS and does not depend on the
ignore.strand
argument.
See the section on ignore.strand
for details.
mapToTranscripts
, pmapToTranscripts
The genomic range in x
is mapped to the local position in the
transcripts
ranges. A successful mapping occurs when x
is completely within the transcripts
range, equivalent to:
findOverlaps(..., type="within")
Transcriptome-based coordinates start counting at 1 at the beginning
of the transcripts
range and return positions where x
was aligned. The seqlevels of the return object are taken from the
transcripts
object and should be transcript names. In this
direction, mapping is attempted between all elements of x
and
all elements of transcripts
.
mapToTranscripts
uses findOverlaps
to map ranges in
x
to ranges in transcripts
. This method does not return
unmapped ranges.
pmapToTranscripts
maps the i-th range in x
to the
i-th range in transcripts
. Recycling is supported for both
x
and transcripts
when either is length == 1L; otherwise
the lengths must match. Ranges in x
that do not map (out of bounds
or strand mismatch) are returned as zero-width ranges starting at 0.
These ranges are given the seqname of "UNMAPPED".
mapFromTranscripts
, pmapFromTranscripts
The transcript-based position in x
is mapped to genomic coordinates
using the ranges in transcripts
. A successful mapping occurs when
the following is TRUE:
width(transcripts) >= start(x) + width(x)
x
is aligned to transcripts
by moving in start(x)
positions in from the beginning of the transcripts
range. The
seqlevels of the return object are chromosome names.
mapFromTranscripts
uses the seqname of x
and the names
of transcripts
to determine mapping pairs (vs attempting to match
all possible pairs). Name matching is motivated by use cases such as
differentially expressed regions where the expressed regions in x
would only be related to a subset of regions in transcripts
.
This method does not return unmapped ranges.
pmapFromTranscripts
maps the i-th range in x
to the i-th
range in transcripts
and therefore does not use name matching.
Recycling is supported in pmapFromTranscripts
when either
x
or transcripts
is length == 1L; otherwise the lengths
must match. Ranges in x
that do not map (out of bounds or strand
mismatch) are returned as zero-width ranges starting at 0. These ranges
are given the seqname of "UNMAPPED".
pmapToTranscripts
returns a GRanges
the same length as
x
.
pmapFromTranscripts
returns a GRanges
when transcripts
is a GRanges
and a GRangesList
when transcripts
is a GRangesList
. In both cases the return object is the same
length as x
. The rational for returning the GRangesList
is
to preserve exon structure; ranges in a list element that are not overlapped
by x
are returned as a zero-width range. The GRangesList
return object will have no seqlevels called "UNMAPPED"; those will only
occur when a GRanges
is returned.
mapToTranscripts
and mapFromTranscripts
return GRanges
objects that vary in length similar to a Hits
object. The result
contains mapped records only; strand mismatch and out of bound ranges are
not returned. xHits
and transcriptsHits
metadata columns
(similar to the queryHits
and subjectHits
of a Hits
object) indicate elements of x
and transcripts
used in
the mapping.
When intronJunctions
is TRUE, mapToTranscripts
returns an
extra metdata column named intronic
to identify the intron ranges.
When mapping to transcript coordinates, seqlevels of the output are the names
on the transcripts
object and most often these will be transcript
names. When mapping to the genome, seqlevels of the output are the seqlevels
of transcripts
which are usually chromosome names.
V. Obenchain, M. Lawrence and H. Pagès
?mapToAlignments
in the
GenomicAlignments package for methods mapping between
reads and genome space using a CIGAR alignment.
## --------------------------------------------------------------------- ## A. Basic Use ## --------------------------------------------------------------------- ## ------------------------------------ ## (i) Map from genome to transcript: ## The seqnames of the output are the transcript names, not chromosomes. For ## this reason 'transcripts' must be named. x <- GRanges("A", IRanges(16, 18)) gr1 <- GRanges("A", IRanges(1, 10, names="tx_a")) gr2 <- GRanges("A", IRanges(15, 20, names="tx_b")) ## 'transcripts' as GRanges: mapToTranscripts(x, gr2) ## 'transcripts' as GRangesList: mapToTranscripts(x, GRangesList("tx_c" = c(gr1, gr2))) ## Round trip from genomic -> transcript -> genomic coordinates: tx_coord <- mapToTranscripts(x, gr2) mapFromTranscripts(tx_coord, gr2) ## ------------------------------------ ## (ii) Map from transcript to genome: ## A prerequisite for mapping from transcript -> genome is that the seqname ## of the range in 'x' match the name of the range in 'transcripts'. Here ## the seqname of 'x' is "TX_1" and mapping is only attempted with the second ## range in 'gr': x <- GRanges("TX_1", IRanges(5, 10)) gr <- GRanges("chr3", IRanges(c(1, 1), width=50, names=c("TX_2", "TX_1"))) mapFromTranscripts(x, gr) ## ------------------------------------ ## (iii) Element-wise versions: ## Recycling is supported when length(transcripts) == 1; otherwise the ## lengths of 'x' and 'transcripts' must be the same. x <- GRanges("A", IRanges(c(1, 5, 10), width=1)) transcripts <- GRanges("A", IRanges(4, 7)) pmapToTranscripts(x, transcripts) ## --------------------------------------------------------------------- ## B. Map local sequence locations to the genome ## --------------------------------------------------------------------- ## NAGNAG alternative splicing plays an essential role in biological processes ## and represents a highly adaptable system for posttranslational regulation ## of gene function. The majority of NAGNAG studies largely focus on messenger ## RNA. A study by Sun, Lin, and Yan ## (http://www.hindawi.com/journals/bmri/2014/736798/) demonstrated that ## NAGNAG splicing is also operative in large intergenic noncoding RNA ## (lincRNA). ## One finding of interest was that linc-POLR3G-10 exhibited two NAGNAG ## acceptors located in two distinct transcripts: TCONS_00010012 and ## TCONS_00010010. ## Extract the exon coordinates of TCONS_00010012 and TCONS_00010010: lincrna <- c("TCONS_00010012", "TCONS_00010010") library(TxDb.Hsapiens.UCSC.hg19.lincRNAsTranscripts) txdb <- TxDb.Hsapiens.UCSC.hg19.lincRNAsTranscripts exons <- exonsBy(txdb, by="tx", use.names=TRUE)[lincrna] exons ## The two NAGNAG acceptors were identified in the upstream region of ## the fourth and fifth exons located in TCONS_00010012. ## Extract the sequences for transcript TCONS_00010012: library(BSgenome.Hsapiens.UCSC.hg19) genome <- BSgenome.Hsapiens.UCSC.hg19 exons_seq <- getSeq(genome, exons[[1]]) ## TCONS_00010012 has 4 exons: exons_seq ## The most common triplet among the lincRNA sequences was CAG. Identify ## the location of this pattern in all exons. cag_loc <- vmatchPattern("CAG", exons_seq) ## Convert the first occurance of CAG in each exon back to genome coordinates. first_loc <- do.call(c, sapply(cag_loc, "[", 1, simplify=TRUE)) pmapFromTranscripts(first_loc, exons[[1]]) ## ----------------------------------------------------------------------- ## C. Map 3'UTR variants to genome coordinates ## ----------------------------------------------------------------------- ## A study by Skeeles et. al (PLoS ONE 8(3): e58609. doi: ## 10.1371/journal.pone.0058609) investigated the impact of 3'UTR variants ## on the expression of cancer susceptibility genes. ## 8 candidate miRNA genes on chromosome 12 were used to test for ## differential luciferase expression in mice. In Table 2 of the manuscript ## variant locations are given as nucleotide position within the gene. geneNames <- c("Bcap29", "Dgkb", "Etv1", "Hbp1", "Hbp1", "Ifrd1", "Ifrd1", "Pik3cg", "Pik3cg", "Tspan13", "Twistnb") starts <- c(1409, 3170, 3132, 2437, 2626, 3239, 3261, 4947, 4979, 958, 1489) snps <- GRanges(geneNames, IRanges(starts, width=1)) ## To map these transcript-space coordinates to the genome we need gene ranges ## in genome space. library(org.Mm.eg.db) geneid <- select(org.Mm.eg.db, unique(geneNames), "ENTREZID", "SYMBOL") geneid ## Extract the gene regions: library(TxDb.Mmusculus.UCSC.mm10.knownGene) txdb <- TxDb.Mmusculus.UCSC.mm10.knownGene genes <- genes(txdb)[geneid$ENTREZID] ## A prerequesite of the mapping from transcript space to genome space ## is that seqnames in 'x' match names in 'transcripts'. Rename ## 'genes' with the appropriate gene symbol. names(genes) <- geneid$SYMBOL ## The xHits and transcriptsHits metadta columns indicate which ranges in ## 'snps' and 'genes' were involved in the mapping. mapFromTranscripts(snps, genes) ## ----------------------------------------------------------------------- ## D. Map dbSNP variants to cds or cDNA coordinates ## ----------------------------------------------------------------------- ## The GIPR gene encodes a G-protein coupled receptor for gastric inhibitory ## polypeptide (GIP). Originally GIP was identified to inhibited gastric acid ## secretion and gastrin release but was later demonstrated to stimulate ## insulin release in the presence of elevated glucose. ## In this example 5 SNPs located in the GIPR gene are mapped to cDNA ## coordinates. A list of SNPs in GIPR can be downloaded from dbSNP or NCBI. rsids <- c("rs4803846", "rs139322374", "rs7250736", "rs7250754", "rs9749185") ## Extract genomic coordinates with a SNPlocs package. library(SNPlocs.Hsapiens.dbSNP144.GRCh38) snps <- snpsById(SNPlocs.Hsapiens.dbSNP144.GRCh38, rsids) ## Gene regions of GIPR can be extracted from a TxDb package of compatible ## build. The TxDb package uses Entrez gene identifiers and GIPR is a gene ## symbol. Conversion between gene symbols and Entrez gene IDs is done by ## calling select() on an organism db package. library(org.Hs.eg.db) geneid <- select(org.Hs.eg.db, "GIPR", "ENTREZID", "SYMBOL") ## The transcriptsBy() extractor returns a range for each transcript that ## includes the UTR and exon regions (i.e., cDNA). library(TxDb.Hsapiens.UCSC.hg38.knownGene) txdb <- TxDb.Hsapiens.UCSC.hg38.knownGene txbygene <- transcriptsBy(txdb, "gene") cDNA <- txbygene[geneid$ENTREZID] cDNA ## Before mapping, the chromosome names (seqlevels) in the two objects must ## be harmonized. The style is NCBI for 'snps' and UCSC for 'cDNA'. seqlevelsStyle(snps) seqlevelsStyle(cDNA) ## Modify the style and genome in 'snps' to match 'cDNA'. seqlevelsStyle(snps) <- seqlevelsStyle(cDNA) genome(snps) <- genome(cDNA) ## The 'cDNA' object is a GRangesList of length 1. This single list element ## contains the cDNA range for 4 different transcripts. To map to each ## transcript individually 'cDNA' must be unlisted before mapping. ## Map all 5 SNPS to all 4 transcripts: mapToTranscripts(snps, unlist(cDNA)) ## Map the first SNP to transcript uc002pct.2 and the second to uc002pcu.2. pmapToTranscripts(snps[1:2], unlist(cDNA)[1:2]) ## The cdsBy() extractor returns coding regions by gene or by transcript. ## Extract the coding regions for transcript uc002pct.2. cds <- cdsBy(txdb, "tx", use.names=TRUE)["uc002pct.2"] cds ## The 'cds' object is a GRangesList of length 1 containing all cds ranges ## for the single transcript uc002pct.2. ## To map to the concatenated group of ranges leave 'cds' as a GRangesList. mapToTranscripts(snps, cds) ## Only the second SNP could be mapped. Unlisting the 'cds' object maps the ## SNPs to the individual cds ranges (vs the concatenated range). mapToTranscripts(snps[2], unlist(cds)) ## The location is the same because the SNP hit the first cds range. If the ## transcript were on the "-" strand the difference in concatenated vs ## non-concatenated position would be more obvious. ## Change strand: strand(cds) <- strand(snps) <- "-" mapToTranscripts(snps[2], unlist(cds))