Aggregate data from used run elements

[mwhamgenomics]

In the sample report, there is a discrepancy between yield (calculated dynamically from usable run elements) and coverage, duplicates, etc. (calculated upon sample processing). The dynamic yield calculation is only needed for the DatasetScanner in the pipeline and is done in the run element endpoint anyway.

We should have the sample report aggregate data only from the run elements in data_source, rather than all run elements marked usable at the time.

Edit: instead of this, we should highlight instances of run element mismatching as mentioned below.

[tcezard]

I'm not sure the analysis_driver_proc.data_source will be any more reliable than taking the available useable run_elements.
A short term solution is to add a visual aid to highlight the discrepancy while it exists.
We can check the data_source and the currently useable run_elements and colour the row in red if they do not match.

[tcezard]

see #112 for more long term solution