Getting QC reports

[davidmosen]

We need to take a look at aw_stats.r and update it to get more stats. We can integrate those with the generated plots to create a report.

[antoniojbt]

What's the purpose of the script? It looks like it currently puts together a list of included and excluded individuals?

Are we looking for relatively simple marker and individuals stats? If so we could reformat/save/rename plink's logs and run it's summary statistics commands.

[davidmosen]

I'm working on something a bit more comprehensive and polished. Will share it soon and we can brainstorm what else would be needed for a barebone report.

[antoniojbt]

Sounds good!

[davidmosen]

Please see "/groupvol/med-bio/epiUKB/Airwave/QC/affy_b37_maf0.01/report/"

You will need to download the folder and then open the file called "report.html" to view it properly.

[antoniojbt]

I had a quick scan, looks excellent! I didn't check the actual results, we should meet/email to discuss these separately.

I haven't seen how you generated it. Is it easy to transfer to other planned pipelines? Easy to learn?
Thanks!

[davidmosen]

Antonio, the report is generated by lines 90-93 in file aw_all_batches.sh:

cd ${PBS_O_WORKDIR}
cp -r assets ${OUTPATH}/report/
Rscript -e "library(knitr); knit('report.Rhtml', output='${OUTPATH}/report/report.html')" $PARAMS

The core code is in report.Rhtml, which works just like Sweave or the markdown for plain text that you were using, with the only difference that this mixes R with HTML, rather than LaTeX or plain text. The background HTML+CSS+JS design can be applied to other pipelines/reports for sure.