expression_filter_fusion: Expression filtering with user provided expression Matrix and...

In d3b-center/annoFuse: annoFuse: an R Package to annotate, prioritize, and interactively explore putative oncogenic RNA fusions

Expression filtering with user provided expression Matrix and standard fusion calls

Description

Expression filtering with user provided expression Matrix and standard fusion calls

Usage

expression_filter_fusion(
  standardFusioncalls,
  expressionMatrix,
  expressionFilter
)

Arguments

standardFusioncalls	A dataframe from star fusion or arriba standardized to run through the filtering steps
expressionMatrix	Expression matrix for samples used in cohort for fusion calls
expressionFilter	FPKM/TPM threshold for not expressed

Value

Standardized fusion calls annotated with gene list and fusion list provided in reference folder

Examples

## Not run: 
# standardize
fusionfileArriba <- read_arriba_calls(
  system.file("extdata", "arriba_example.tsv", package = "annoFuseData")
)
fusionfileStarFusion <- read_starfusion_calls(
  system.file("extdata", "starfusion_example.tsv", package = "annoFuseData")
)
library(dplyr)
formattedArriba <- fusion_standardization(fusionfileArriba,
  caller = "ARRIBA",
  tumorID = "tumorID"
)
formattedStarFusion <- fusion_standardization(fusionfileStarFusion,
  caller = "STARFUSION",
  tumorID = "tumorID"
)
# merge standardized fusion calls
standardFusioncalls <- rbind(formattedStarFusion, formattedArriba) %>% as.data.frame()
fusionQCFiltered <- fusion_filtering_QC(
  standardFusioncalls = standardFusioncalls,
  readingFrameFilter = "in-frame|frameshift|other",
  artifactFilter = "GTEx_Recurrent|DGD_PARALOGS|Normal|BodyMap|ConjoinG",
  junctionReadCountFilter = 1,
  spanningFragCountFilter = 10,
  readthroughFilter = TRUE
)
# expression based filter to capture only fusions where atleast 1 gene is expressed

expressionFile <- system.file("extdata", "example.rsem.genes.results.gz", package = "annoFuseData")
expressionMatrix <- read_tsv(expressionFile)
library(reshape2)
# split gene id and symbol
expressionMatrix <- cbind(
  expressionMatrix,
  colsplit(expressionMatrix$gene_id, pattern = "_", names = c("EnsembleID", "GeneSymbol"))
)
# collapse to matrix of HUGO symbols x Sample identifiers
# take max expression per row and use the max value for duplicated gene symbols
expressionMatrix.collapsed <- expressionMatrix %>%
  arrange(desc(FPKM)) %>% # arrange decreasing by FPKM
  distinct(GeneSymbol, .keep_all = TRUE) %>% # keep the ones with greatest FPKM value.
  # If ties occur, keep the first occurencce
  unique() %>%
  remove_rownames() %>%
  dplyr::select(.data$EnsembleID, .data$GeneSymbol, .data$FPKM, .data$gene_id)
# rename columns
colnames(expressionMatrix.collapsed)[3] <- "tumorID"
expressionFiltered <- expression_filter_fusion(
  standardFusioncalls = fusionQCFiltered,
  expressionMatrix = expressionMatrix.collapsed,
  expressionFilter = 1
)

## End(Not run)

d3b-center/annoFuse documentation built on Oct. 2, 2024, 4:17 a.m.