If you are certain that coincidental emission are not the culprits, that
only leaves one other thing in my mind. Unfortunately, it will be harder
to track this one down than the others.
As you probably know, the Optima analyzes the entire array under MSF
instead of just looking at the few pixels near the peak. This means you
can have an emission quite a distance from your analyte's emission
included in the quantitation. Conversely, if you have a sample with a
significantly different matrix from what you have defined as your
'blank' spectra, you can also come up with false negatives.
One way to minimize this effect is to judiciously use their masking
feature when building your MSF file, which blocks off portions of the
array from being measured.
As for PE support, may I suggest you contact the 'guru' Mike Duffy here
in the states. It would be worth a call. He's awfully busy, but pretty
good about answering his messages. You can also email him at
<[log in to unmask]>.
Best of luck. It sounds like a real teaser.