Run 140828_M01841_0073_000000000-A8CHE had poor data from one surface

[cbrumme]

The TOP surface started to give poor data at around cycle 279: At that point we start to see loss of focus in tiles 1101-1104 every 6 cycles. This can be observed in a plot of Error Rate vs. Cycle or FWHM vs Cycle, by tile.

We noticed the error because there were many nucleotide mismatches in the overlapping sections of R1 and R2. This caused the run to fail HLA interpretation. It's unknown whether this type of error would have affected the HIV workflow.

At the moment we don't know how this error would reflect in the run-average metrics on the QC report as this run is one of those affected by the Oracle uploading bugs.

If the QC report is unable to identify such an error this might be a useful run to use in designing more in-depth metrics