EnsDb-exonsBy: Retrieve annotation data from an Ensembl based package

In YTLogos/ensembldb: Utilities to create and use Ensembl-based annotation databases

Description

Retrieve gene/transcript/exons annotations stored in an Ensembl based database package generated with the makeEnsembldbPackage function. Parameter filter enables to define filters to retrieve only specific data. Alternatively, a global filter might be added to the EnsDb object using the addFilter method.

Usage

## S4 method for signature 'EnsDb'
exons(x, columns = listColumns(x,"exon"),
        filter = AnnotationFilterList(), order.by,
        order.type = "asc", return.type = "GRanges")

## S4 method for signature 'EnsDb'
exonsBy(x, by = c("tx", "gene"),
        columns = listColumns(x, "exon"), filter =
        AnnotationFilterList(), use.names = FALSE)

## S4 method for signature 'EnsDb'
exonsByOverlaps(x, ranges, maxgap = 0L, minoverlap = 1L,
        type = c("any", "start", "end"), columns = listColumns(x, "exon"),
        filter = AnnotationFilterList())

## S4 method for signature 'EnsDb'
transcripts(x, columns = listColumns(x, "tx"),
        filter = AnnotationFilterList(), order.by, order.type = "asc",
        return.type = "GRanges")

## S4 method for signature 'EnsDb'
transcriptsBy(x, by = c("gene", "exon"),
        columns = listColumns(x, "tx"), filter = AnnotationFilterList())

## S4 method for signature 'EnsDb'
transcriptsByOverlaps(x, ranges, maxgap = 0L,
        minoverlap = 1L, type = c("any", "start", "end"),
        columns = listColumns(x, "tx"), filter = AnnotationFilterList())

## S4 method for signature 'EnsDb'
promoters(x, upstream = 2000, downstream = 200, ...)

## S4 method for signature 'EnsDb'
genes(x, columns = c(listColumns(x, "gene"), "entrezid"),
        filter = AnnotationFilterList(), order.by, order.type = "asc",
        return.type = "GRanges")

## S4 method for signature 'EnsDb'
disjointExons(x, aggregateGenes = FALSE,
        includeTranscripts = TRUE, filter = AnnotationFilterList(), ...)

## S4 method for signature 'EnsDb'
cdsBy(x, by = c("tx", "gene"), columns = NULL,
        filter = AnnotationFilterList(), use.names = FALSE)

## S4 method for signature 'EnsDb'
fiveUTRsByTranscript(x, columns = NULL,
        filter = AnnotationFilterList())

## S4 method for signature 'EnsDb'
threeUTRsByTranscript(x, columns = NULL,
        filter = AnnotationFilterList())

## S4 method for signature 'GRangesList'
toSAF(x, ...)

Arguments

(In alphabetic order)

...	For promoters: additional arguments to be passed to the transcripts method.
aggregateGenes	For disjointExons: When FALSE (default) exon fragments that overlap multiple genes are dropped. When TRUE, all fragments are kept and the gene_id metadata column includes all gene IDs that overlap the exon fragment.
by	For exonsBy: wheter exons sould be fetched by genes or by transcripts; as in the corresponding function of the GenomicFeatures package. For transcriptsBy: whether transcripts should be fetched by genes or by exons; fetching transcripts by cds as supported by the transcriptsBy method in the GenomicFeatures package is currently not implemented. For cdsBy: whether cds should be fetched by transcript of by gene.
columns	Columns to be retrieved from the database tables. Default values for genes are all columns from the gene database table, for exons and exonsBy the column names of the exon database table table and for transcript and transcriptBy the columns of the tx data base table (see details below for more information). Note that any of the column names of the database tables can be submitted to any of the methods (use listTables or listColumns methods for a complete list of allowed column names). For cdsBy: this argument is only supported for for by="tx".
downstream	For method promoters: the number of nucleotides downstream of the transcription start site that should be included in the promoter region.
filter	A filter describing which results to retrieve from the database. Can be a single object extending AnnotationFilter, an AnnotationFilterList object combining several such objects or a formula representing a filter expression (see examples below or AnnotationFilter for more details). Use the supportedFilters method to get an overview of supported filter classes and related fields.
includeTranscripts	For disjointExons: When TRUE (default) a tx_name metadata column is included that lists all transcript IDs that overlap the exon fragment. Note: this is different to the disjointExons function in the GenomicFeatures package, that lists the transcript names, not IDs.
maxgap	For exonsByOverlaps and transcriptsByOverlaps: see exonsByOverlaps help page in the GenomicFeatures package.
minoverlap	For exonsByOverlaps and transcriptsByOverlaps: see exonsByOverlaps help page in the GenomicFeatures package.
order.by	Character vector specifying the column(s) by which the result should be ordered. This can be either in the form of "gene_id, seq_name" or c("gene_id", "seq_name").
order.type	If the results should be ordered ascending (asc, default) or descending (desc).
ranges	For exonsByOverlaps and transcriptsByOverlaps: a GRanges object specifying the genomic regions.
return.type	Type of the returned object. Can be either "data.frame", "DataFrame" or "GRanges". In the latter case the return object will be a GRanges object with the GRanges specifying the chromosomal start and end coordinates of the feature (gene, transcript or exon, depending whether genes, transcripts or exons was called). All additional columns are added as metadata columns to the GRanges object.
type	For exonsByOverlaps and transcriptsByOverlaps: see exonsByOverlaps help page in the GenomicFeatures package.
upstream	For method promoters: the number of nucleotides upstream of the transcription start site that should be included in the promoter region.
use.names	For cdsBy and exonsBy: only for by="gene": use the names of the genes instead of their IDs as names of the resulting GRangesList.
x	For toSAF a GRangesList object. For all other methods an EnsDb instance.

Details

A detailed description of all database tables and the associated attributes/column names is also given in the vignette of this package. An overview of the columns is given below:

gene_id

the Ensembl gene ID of the gene.

gene_name

the name of the gene (in most cases its official symbol).

entrezid

the NCBI Entrezgene ID of the gene. Note that this column contains a list of Entrezgene identifiers to accommodate the potential 1:n mapping between Ensembl genes and Entrezgene IDs.

gene_biotype

the biotype of the gene.

gene_seq_start

the start coordinate of the gene on the sequence (usually a chromosome).

gene_seq_end

the end coordinate of the gene.

seq_name

the name of the sequence the gene is encoded (usually a chromosome).

seq_strand

the strand on which the gene is encoded

seq_coord_system

the coordinate system of the sequence.

tx_id

the Ensembl transcript ID.

tx_biotype

the biotype of the transcript.

tx_seq_start

the chromosomal start coordinate of the transcript.

tx_seq_end

the chromosomal end coordinate of the transcript.

tx_cds_seq_start

the start coordinate of the coding region of the transcript (NULL for non-coding transcripts).

tx_cds_seq_end

the end coordinate of the coding region.

exon_id

the ID of the exon. In Ensembl, each exon specified by a unique chromosomal start and end position has its own ID. Thus, the same exon might be part of several transcripts.

exon_seq_start

the chromosomal start coordinate of the exon.

exon_seq_end

the chromosomal end coordinate of the exon.

exon_idx

the index of the exon in the transcript model. As noted above, an exon can be part of several transcripts and thus its position inside these transcript might differ.

Many EnsDb databases provide also protein related annotations. See listProteinColumns for more information.

Value

For exons, transcripts and genes, a data.frame, DataFrame or a GRanges, depending on the value of the return.type parameter. The result is ordered as specified by the parameter order.by or, if not provided, by seq_name and chromosomal start coordinate, but NOT by any ordering of values in eventually submitted filter objects.

For exonsBy, transcriptsBy: a GRangesList, depending on the value of the return.type parameter. The results are ordered by the value of the by parameter.

For exonsByOverlaps and transcriptsByOverlaps: a GRanges with the exons or transcripts overlapping the specified regions.

For toSAF: a data.frame with column names "GeneID" (the group name from the GRangesList, i.e. the ID by which the GRanges are split), "Chr" (the seqnames from the GRanges), "Start" (the start coordinate), "End" (the end coordinate) and "Strand" (the strand).

For disjointExons: a GRanges of non-overlapping exon parts.

For cdsBy: a GRangesList with GRanges per either transcript or exon specifying the start and end coordinates of the coding region of the transcript or gene.

For fiveUTRsByTranscript: a GRangesList with GRanges for each protein coding transcript representing the start and end coordinates of full or partial exons that constitute the 5' untranslated region of the transcript.

For threeUTRsByTranscript: a GRangesList with GRanges for each protein coding transcript representing the start and end coordinates of full or partial exons that constitute the 3' untranslated region of the transcript.

Methods and Functions

exons

Retrieve exon information from the database. Additional columns from transcripts or genes associated with the exons can be specified and are added to the respective exon annotation.

exonsBy

Retrieve exons grouped by transcript or by gene. This function returns a GRangesList as does the analogous function in the GenomicFeatures package. Using the columns parameter it is possible to determine which additional values should be retrieved from the database. These will be included in the GRanges object for the exons as metadata columns. The exons in the inner GRanges are ordered by the exon index within the transcript (if by="tx"), or increasingly by the chromosomal start position of the exon or decreasingly by the chromosomal end position of the exon depending whether the gene is encoded on the + or - strand (for by="gene"). The GRanges in the GRangesList will be ordered by the name of the gene or transcript.

exonsByOverlaps

Retrieve exons overlapping specified genomic ranges. For more information see exonsByOverlaps method in the GenomicFeatures package. The functionality is to some extent similar and redundant to the exons method in combination with GRangesFilter filter.

transcripts

Retrieve transcript information from the database. Additional columns from genes or exons associated with the transcripts can be specified and are added to the respective transcript annotation.

transcriptsBy

Retrieve transcripts grouped by gene or exon. This function returns a GRangesList as does the analogous function in the GenomicFeatures package. Using the columns parameter it is possible to determine which additional values should be retrieved from the database. These will be included in the GRanges object for the transcripts as metadata columns. The transcripts in the inner GRanges are ordered increasingly by the chromosomal start position of the transcript for genes encoded on the + strand and in a decreasing manner by the chromosomal end position of the transcript for genes encoded on the - strand. The GRanges in the GRangesList will be ordered by the name of the gene or exon.

transcriptsByOverlaps

Retrieve transcripts overlapping specified genomic ranges. For more information see transcriptsByOverlaps method in the GenomicFeatures package. The functionality is to some extent similar and redundant to the transcripts method in combination with GRangesFilter filter.

promoters

Retrieve promoter information from the database. Additional columns from genes or exons associated with the promoters can be specified and are added to the respective promoter annotation.

genes

Retrieve gene information from the database. Additional columns from transcripts or exons associated with the genes can be specified and are added to the respective gene annotation. Note that column "entrezid" is a list of Entrezgene identifiers to accomodate the potential 1:n mapping between Ensembl genes and Entrezgene IDs.

disjointExons

This method is identical to disjointExons defined in the GenomicFeatures package. It creates a GRanges of non-overlapping exon parts with metadata columns of gene_id and exonic_part. Exon parts that overlap more than one gene can be dropped with aggregateGenes=FALSE.

cdsBy

Returns the coding region grouped either by transcript or by gene. Each element in the GRangesList represents the cds for one transcript or gene, with the individual ranges corresponding to the coding part of its exons. For by="tx" additional annotation columns can be added to the individual GRanges (in addition to the default columns exon_id and exon_rank). Note that the GRangesList is sorted by its names.

fiveUTRsByTranscript

Returns the 5' untranslated region for protein coding transcripts.

threeUTRsByTranscript

Returns the 3' untranslated region for protein coding transcripts.

toSAF

Reformats a GRangesList object into a data.frame corresponding to a standard SAF (Simplified Annotation Format) file (i.e. with column names "GeneID", "Chr", "Start", "End" and "Strand"). Note: this method makes only sense on a GRangesList that groups features (exons, transcripts) by gene.

Note

Ensembl defines genes not only on standard chromosomes, but also on patched chromosomes and chromosome variants. Thus it might be advisable to restrict the queries to just those chromosomes of interest (e.g. by specifying a SeqNameFilter(c(1:22, "X", "Y"))). In addition, also so called LRG genes (Locus Reference Genomic) are defined in Ensembl. Their gene id starts with LRG instead of ENS for Ensembl genes, thus, a filter can be applied to specifically select those genes or exclude those genes (see examples below).

Depending on the value of the global option "ucscChromosomeNames" (use getOption(ucscChromosomeNames, FALSE) to get its value or option(ucscChromosomeNames=TRUE) to change its value) the sequence/chromosome names of the returned GRanges objects or provided in the returned data.frame or DataFrame correspond to Ensembl chromosome names (if value is FALSE) or UCSC chromosome names (if TRUE). This ensures a better integration with the Gviz package, in which this option is set by default to TRUE.

Note

While it is possible to request values from a column "tx_name" (with the columns argument), no such column is present in the database. The returned values correspond to the ID of the transcripts.

Author(s)

Johannes Rainer, Tim Triche

See Also

supportedFilters to get an overview of supported filters. makeEnsembldbPackage, listColumns, lengthOf

addFilter for globally adding filters to an EnsDb object.

Examples

library(EnsDb.Hsapiens.v75)
edb <- EnsDb.Hsapiens.v75

######   genes
##
## Get all genes encoded on chromosome Y
AllY <- genes(edb, filter = SeqNameFilter("Y"))
AllY

## Return the result as a DataFrame; also, we use a filter expression here
## to define which features to extract from the database.
AllY.granges <- genes(edb,
                      filter = ~ seq_name == "Y",
                      return.type="DataFrame")
AllY.granges

## Include all transcripts of the gene and their chromosomal
## coordinates, sort by chrom start of transcripts and return as
## GRanges.
AllY.granges.tx <- genes(edb,
                         filter = SeqNameFilter("Y"),
                         columns = c("gene_id", "seq_name",
                                     "seq_strand", "tx_id", "tx_biotype",
                                     "tx_seq_start", "tx_seq_end"),
                         order.by = "tx_seq_start")
AllY.granges.tx



######   transcripts
##
## Get all transcripts of a gene
Tx <- transcripts(edb,
                  filter = GeneIdFilter("ENSG00000184895"),
                  order.by = "tx_seq_start")
Tx

## Get all transcripts of two genes along with some information on the
## gene and transcript
Tx <- transcripts(edb,
                  filter = GeneIdFilter(c("ENSG00000184895",
                                          "ENSG00000092377")),
                  columns = c("gene_id", "gene_seq_start", "gene_seq_end",
                              "gene_biotype", "tx_biotype"))
Tx

######   promoters
##
## Get the bona-fide promoters (2k up- to 200nt downstream of TSS)
promoters(edb, filter = GeneIdFilter(c("ENSG00000184895",
                                       "ENSG00000092377")))

######   exons
##
## Get all exons of protein coding transcript for the gene ENSG00000184895
Exon <- exons(edb,
              filter = ~ gene_id == "ENSG00000184895" &
                  tx_biotype == "protein_coding",
              columns = c("gene_id", "gene_seq_start", "gene_seq_end",
                          "tx_biotype", "gene_biotype"))
Exon



#####    exonsBy
##
## Get all exons for transcripts encoded on chromosomes X and Y.
ETx <- exonsBy(edb, by = "tx",
               filter = SeqNameFilter(c("X", "Y")))
ETx
## Get all exons for genes encoded on chromosome 1 to 22, X and Y and
## include additional annotation columns in the result
EGenes <- exonsBy(edb, by = "gene",
                  filter = SeqNameFilter(c("X", "Y")),
                  columns = c("gene_biotype", "gene_name"))
EGenes

## Note that this might also contain "LRG" genes.
length(grep(names(EGenes), pattern="LRG"))

## to fetch just Ensemblgenes, use an GeneIdFilter with value
## "ENS%" and condition "like"
eg <- exonsBy(edb, by = "gene",
              filter = AnnotationFilterList(SeqNameFilter(c("X", "Y")),
                                            GeneIdFilter("ENS", "startsWith")),
              columns = c("gene_biotype", "gene_name"))
eg
length(grep(names(eg), pattern="LRG"))

#####    transcriptsBy
##
TGenes <- transcriptsBy(edb, by = "gene",
                        filter = SeqNameFilter(c("X", "Y")))
TGenes

## convert this to a SAF formatted data.frame that can be used by the
## featureCounts function from the Rsubreader package.
head(toSAF(TGenes))


#####   transcriptsByOverlaps
##
ir <- IRanges(start = c(2654890, 2709520, 28111770),
              end = c(2654900, 2709550, 28111790))
gr <- GRanges(rep("Y", length(ir)), ir)

## Retrieve all transcripts overlapping any of the regions.
txs <- transcriptsByOverlaps(edb, gr)
txs

## Alternatively, use a GRangesFilter
grf <- GRangesFilter(gr, type = "any")
txs <- transcripts(edb, filter = grf)
txs


####    cdsBy
## Get the coding region for all transcripts on chromosome Y.
## Specifying also additional annotation columns (in addition to the default
## exon_id and exon_rank).
cds <- cdsBy(edb, by = "tx", filter = SeqNameFilter("Y"),
             columns = c("tx_biotype", "gene_name"))

####    the 5' untranslated regions:
fUTRs <- fiveUTRsByTranscript(edb, filter = SeqNameFilter("Y"))

####    the 3' untranslated regions with additional column gene_name.
tUTRs <- threeUTRsByTranscript(edb, filter = SeqNameFilter("Y"),
                               columns = "gene_name")

YTLogos/ensembldb documentation built on May 3, 2019, 9:03 p.m.