Still seeing unexpected behavior with 1.0-1

[Matt Shirley]

I am still seeing a discrepancy between the number of samples with CNV reported while visualizing a specific region. Here is some detailed output that shows the behavior:

Code:

filename = paste("chr",chrom,":",start,"-",end,".pdf",sep="")

pdf(file = filename)

visualizeRegion(CNVdata =

objCNVineta,batch.case=c("Cell"),batch.control=c("Whole"), chrom = chrom ,

start = start, end = end,add.heatmap=TRUE)

dev.off()

R output:

 3849  samples to process

2  samples with CNV

[1] "cases   : 1 segments in  1 samples"

[1] "controls: 1 segments in  1 samples"

[1] "got 0 transcripts"

3849  samples to process

2  samples with CNV

     size 0 1  2 3 4+ #del samples #dup samples #CNP samples

Cell 1335 0 0 NA 0  0            0            0            0

sum  1335 0 0 NA 0  0            0            0            0

      size 0 1  2 3 4+ #del samples #dup samples #CNP samples

Whole 2514 0 0 NA 1  0            0            1            1

sum   2514 0 0 NA 1  0            0            1            1

                      Deletion Duplication Copy Number Polymorphism

Pearson's Chi-squared        0   0.7475869                0.7475869

Fisher's Exact               1   1.0000000                1.0000000

PDF output 

It is my expectation that the number of CNV should at all times be for the region that I specified. I understand that the graphics have not been updated, and that it may be very difficult to change the behavior of the plotting code, but there should be some explicit warning that the number of CNV reported in the first part of the R output ("X samples with CNV") and the graphical output ("X/Y cases/controls with CNVs") are not for the region specified. Thank you though, for making the tabular output and statistics accurate for the region specified.

 

[Matt Shirley]

Gah! I just read your response to our previous conversation where you said you are aware of this behavior. Sorry for the spam, but I guess it shows that this is still confusing behavior. 

[Michael]

You are right. Some parts of the output is just tracing which I really
should disable.
e.g.:

> 3849 samples to process
>2 samples with CNV
>[1] "cases : 1 segments in 1 samples"
>[1] "controls: 1 segments in 1 samples"
>[1] "got 0 transcripts"

BTW, in the getHeatmap function is an additional tracing output which
sometimes returns a FALSE.
I put these two points on the CNVineta to-do list. With the next
package this will be disabled ...