In Marlin-Na/TnT: Interactive Visualization for Genomic Features

A TxTrack displays transcripts and their stucture (exons, coding sequence) along genomic coordinate. It can be constructed from either a TxDb object by TxTrackFromTxDb or a GRanges object by TxTrackFromGRanges.

For the constructing method from GRanges, two necessary meta-columns ("type", "tx_id") of the GRanges are required. The "tx_id" indicates grouping, and the "type" can be "exon" or "cds" by which ranges of "cds" will be filled with color. This method can be used together with biovizBase::crunch to fetch gene model in a certain region or of a certain gene from TxDb or EnsDb.

The constructed track may be further modified to adjust color, tooltip, display labels, etc.

Similar to the relationship between GeneTrack and FeatureTrack, when the display method of TxTrack applied to genomic feature that may have gaps on genomic coordinate, e.g. RNA-related features that are mapped to genomic coordinate, it is called GroupFeatureTrack. It can be constructed from a GRangesList object by GroupFeatureTrack function, assuming ranges in each group are on the same strand and do not overlap.

suppressPackageStartupMessages({
    library(TnT)
    library(GenomicFeatures)
    library(TxDb.Hsapiens.UCSC.hg19.knownGene)
    library(EnsDb.Hsapiens.v75)
})
txdb <- TxDb.Hsapiens.UCSC.hg19.knownGene

Transcript Track From TxDb

txtrack <- TnT::TxTrackFromTxDb(txdb, height = 400)
TnTGenome(txtrack, view.range = GRanges("chr17", IRanges(41190000, 41290000)))

Please note that TxTrackFromTxDb currently consumes a relative large amount of memory, currently you can use seqlevel argument to limit the chromosomes to extract data from, e.g. TxTrackFromTxDb(txdb, seqlevel = "chr17"). But in future, we will use S4 class CompressedSplitDataFrameList to store the exons and implement direct conversion method to JSON to avoid the intermediate structure as list of data frame.

Transcript Track From GRanges (Fetched From EnsDb)

We can use function crunch from biovizBase package to fetch gene model within a certain genomic region or of a certain gene.

gr <- biovizBase::crunch(EnsDb.Hsapiens.v75, ~ symbol == "BRCA2")
gr

Then we use this GRanges object to construct the TxTrack.

brca2tx <- TnT::TxTrackFromGRanges(gr, label = "BRCA2, transcripts", color = "grey2", height = 300)
TnTGenome(brca2tx)

Let's compare it with output of ggbio!

library(ggbio)
ggplot() + geom_alignment(EnsDb.Hsapiens.v75, which = ~ symbol == "BRCA2")

Modify Track Color and Tooltip

Following the example above, we can fill the tooltip with more information and change the colors.

brca2tx$tooltip <- select(EnsDb.Hsapiens.v75,
    keys = brca2tx$tooltip$tx_id, keytype = "TXID", columns = c("GENEID", "SYMBOL", "TXBIOTYPE")
)
brca2tx$color         <- TnT::mapcol(brca2tx$tooltip$TXBIOTYPE)
brca2tx$display_label <- TnT::strandlabel(
    paste(brca2tx$tooltip$SYMBOL, brca2tx$tooltip$TXBIOTYPE), strand(TnT::trackData(brca2tx))
)
TnTGenome(brca2tx)

GroupFeatureTrack From GRangesList

grl <- GRangesList(
    GRanges("chr12", IRanges(c(100,  200, 1000), width = 50), strand = "+"),
    GRanges("chr12", IRanges(c(900, 1300, 1400), width = 50), strand = "-"),
    GRanges("chr12", IRanges(c(900, 1300, 1400), width = 50), strand = "-"),
    GRanges("chr12", IRanges(c(900, 1300, 1400), width = 50), strand = "*"),
    GRanges("chr12", IRanges(c(900, 1300, 1400), width = 50), strand = "-"),
    GRanges("chr12", IRanges(c(900, 1300, 1400), width = 50), strand = "-"),
    GRanges()
)
names(grl) <- LETTERS[1:7]
grltrack <- TnT::GroupFeatureTrack(grl, color = "steelblue")
TnTGenome(grltrack)