fusionTable: Determine all possible fusions of a Rearrangement object

In cancer-genomics/trellis: Somatic structural variant analysis

Determine all possible fusions of a Rearrangement object

Description

Determine all possible fusions of a Rearrangement object and evaluate whether each is in-frame.

Usage

fusionTable(robj, txdb, tx, cds, genome, orgdb, id = "")

Arguments

robj	A Rearrangement object
txdb	A TxDb object
tx	A GRanges object of transcripts from the TxDb object
cds	A GRangesList object containing the CDS for each transcript
genome	A BSgenome object
orgdb	A OrdDb object for the Hsapiens genome
id	A length-one character vector of the sample identifier

Details

Fusions are analyzed as follows. First, all CDS from genes overlapping a rearrangement are extracted (ignoring strand) using the getCDS function. In addition to extracting all the CDS, this function returns the possible gene fusion that may result from a rearrangement based on the modal rearrangement type (the modal rearrangement type is inferred from the strand and position orientation of read pairs). The genes are denoted generically by

            A          C
 5+ --------------|---------
 3- --------B-----|----D----

where "|" denotes a new sequence junction in a rearranged genome.
  

For a given fusion (say AC), we then clip CDS from A and CDS from B that are absent in the fused product. After clipping, we fuse the remaining CDS from genes A and C. The function tumorProtein is used to derive the amino acid sequence of the tumor protein – the protein that would be formed by fusing the tripped CDS from genes A and C. To assess whether the fusion is in frame, we extract all known full transcripts from genes A and C and translate the DNA sequence of each transcript to an amino acid sequence. We refer to the amino acid sequences of the full CDS as the reference protein. The function referenceProtein is a wrapper for getting the reference amino acid sequences. Given the amino acid sequence of the clipped and fused transcripts (fused tumor protein) and the amino acid sequence of the full, unclipped transcripts (reference protein), we compare their sequences to assess whether the fusion is in-frame using the function inFrameFusions. The results are summarized in tabular format by the function .fusionTable.

See Also

See getCDS for how the CDS from genes involved in a rearrangement are extracted, clip and fuse for how transcripts are clipped and then fused, respectively. See referenceProtein and tumorProtein for deriving germline (unrearranged) and somatic (rearranged) amino acid sequences.

Examples

library(org.Hs.eg.db)
library(BSgenome.Hsapiens.UCSC.hg19)
library(TxDb.Hsapiens.UCSC.hg19.refGene)
txdb <- TxDb.Hsapiens.UCSC.hg19.refGene
genome <- BSgenome.Hsapiens.UCSC.hg19
tx <- transcripts(txdb)
options(warn=-1)
cds.all <- cdsBy(txdb, "tx", use.names=TRUE)
data(rear_list)
r <- rear_list[["18557-18736"]]
fusionTable(r, txdb, tx, cds.all, genome,
            org.Hs.eg.db, id="test")
## in-frame fusion of CTNND2 and TRIO

cancer-genomics/trellis documentation built on Aug. 20, 2024, 5:48 p.m.