[Bioc-devel] ShortRead: optional custom labeling of samples in QA report

[Julian Gehring]

Hi,

In the QA report of the 'ShortRead' package, a short sequential integer 
labeling for referencing the samples/files throughout the report is 
created by default.  Would it be reasonable/possible to allow for other 
optional names to label the samples to make the results of the report 
easier to understand?

In general, I have three ideas what would be handy to have:

1. Derive a label from the file names.  This is probably hard to 
generalize and implement in a way that it actually helps.

2. In case the 'dirPath' argument in the 'qa' function call is a named 
vector, such as

     qa(dirPath=c(p1="bam_file1.bam", p2="bam_file2.bam"))

use the names ["p1", "p2"] for the labeling later on.  This would 
require storing the names in the object returned by 'qa', but should not 
be too hard to implement.

3. Optionally, pass a named vector to the 'report' method, matching file 
names to sample labels.  In case the file names do not match or 
'samples' is missing, default to the sequential labeling.


For option 3, I have created a simple example patch to illustrate how 
this could be implemented (see attached).  So, later this may look like 
this:


     library(ShortRead)
     files = c(p1="bam_file1.bam", p2="bam_file2.bam")
     qa = qa(files, type="BAM")

     ## default sequential labeling ##
     ShortRead:::.report_html_BAMQA(qa, dest="report_normal")

     ## samples named according to names(files) ##
     ShortRead:::.report_html_BAMQA(qa, dest="report_named", samples=files)


I'm happy about any inputs or thoughts regarding this.


Best wishes
Julian