negative T cell fration?

[stroke1989]

Hi, thanks for this excellent tool. I tried to use TcellExTRECT to analyze my WES data. The capture kit I used is IDT version2, not Agilent, so I used TCRA_exons_hg19_custom <- createExonDFBed(PATH_TO_IDT, 'hg19') to create capture kit data frame. However, the final results showed negative T cell fration for normal(blood was used as normal pair in my WES analysis) paired sample bam file.

library(TcellExTRECT)
data("tcra_seg_hg19")
bed.file <- "/home/ug0416/mydata/genome/hg19/hg19_bed/IDTV2.raw.bed"
TCRA_exons_hg19_IDT <- createExonDFBed(bed.file, 'hg19')
mydata <- "~/mydata/05.gatk/a.BQSR/2nd_BQSR/SCA_111_N.sorted.markdup.recal.bam"
cov.file <- getCovFromBam(bamPath = mydata,outPath = '',vdj.seg = tcra_seg_hg19)
cov_df <- loadCov(cov.file)
TCRA.out <- runTcellExTRECT(cov_df, TCRA_exons_hg19_IDT, tcra_seg_hg19, 'hg19')
TCRA.out
sample TCRA.tcell.fraction TCRA.tcell.fraction.lwr TCRA.tcell.fraction.upr qcFit
1 test -1.085798e-13 -86.77173 0.9886068 2.330639
Hope your response! Appreciate!

[rbentham]

Hi,

Please see this related issue: #11

These negative values are a known issue and should be treated as zero infiltration and are due to the handling of floating points in R. The next update will be resolving this and setting these automatically to 0.

There does seem to be a lot of noise in your test sample with the very large lower and upper bounds. I know IDT capture kits can have poor coverage in the TCRA locus. Have you looked at the output of plotTcellExTRECT to see visually what this sample looks like?

[stroke1989]

question_data.pdf
this is the plot, I don't know how to interpret this plot, and it looks so weird. Could you tell me what problem this plot indicate and how can I handle this problem?This T cell fraction was extracted from normal paired sample(blood in our case), so the T cell fraction is abosultely not zero. Is IDT capture kit the key problem? Hope your response! Appreciate!

[squigzzz]

from my understanding, I don't think IDT v2 capture kit covers the TCRA locus thoroughly enough. There seem to be 3-4 genes that are actually sequenced in this region if you look into the bed file.