In RamsinghLab/arkas_staging: A package that complements Kallisto for quick, informative *seq analysis

Introduction Kallisto

Arkas integrates kallisto into the R environment, so kallisto must be installed prior to Arkas calls. Arkas can also be used downstream of kallisto data production specifically for model comparisons.

library(arkas)
library(TxDbLite)


pathBase <- system.file("extdata", "", package="arkasData")
fastaPath <- paste0(pathBase, "/fasta")
fastqPath <- paste0(pathBase, "/fastq/demoFastqDir/")
samples <- c(MrN="MrN", MrT="MrT") ## normally set by appSession
fastaFiles <- c( "ERCC.fa.gz", ## spike-in controls  
                 "Homo_sapiens.RepBase.20_05.merged.fa.gz")

## build an index if it isn't already there (in artemisData, it is)
indexName <- indexKallisto(fastaFiles=fastaFiles, fastaPath=fastaPath,makeUnique=TRUE)$indexName

## run pseudoalignments 
library(parallel)
results <- mclapply(samples, 
                    runKallisto,
                    indexName=indexName,
                    fastqPath=fastqPath,
                    fastaPath=fastaPath,
                    bootstraps=100,
                    outputPath="/data/output")

Heatmaps of Repetitive Elements

Heatmaps can be used to plot repeat elements. Here we plot the top 25 repeats with the highest standard deviation

suppressPackageStartupMessages(library(arkas))

suppressPackageStartupMessages(library(TxDbLite))
samples<-c("n1","n2","n4","s1","s2","s4")
pathBase<-system.file("extdata",package="arkasData")
merged <- mergeKallisto(samples, outputPath=pathBase)
assays(merged)
kallistoVersion(merged)
transcriptomes(merged)
tail(tpm(merged))


## plot repeat element txn using (counts/bootstrap MADs) as "effect size"

bySd <- function(x, k=25) {
   sds<-vector()
  for(i in 1:nrow(x)){
   sds[i]<-sd(x[i,],na.rm=TRUE)
   }
  x[rev(order(sds))[seq_len(k)],]
}
inX<-tpm(merged)
ComplexHeatmap::Heatmap(log(1+bySd(inX[213665:213782,])),
        column_title="Repeat transcription, teratoma vs. normal")

RamsinghLab/arkas_staging documentation built on March 14, 2021, 11:40 a.m.