After doing a couple of experiments with newly immobilized chip, the sensorgram began to give dissociation curve that go below baseline. The problem also that doesn't occur for all the concentrations and it is not dose dependent.
I am using double referencing with the used buffer and I am not using DMSO in my sample or buffer, anyone experienced something like that and what could be the reason?
It must somehowe be ligand related because you observe this only during the dissociation (right?). Does it persist after an overnight standby wash? And when you repeat a concentration series are the response levels comparable with the previous series?
I read the troubleshooting page, however, my case is somewhat different from any mentioned cases.
first, when I have done the same experiment with the same analyte and same chip with the same immobilized Ligand and got nice sensorgram at the first day.
On the day after when I tried to repeat the experiment, I got what I described, it appears more with higher concentration.
Yes it appears only during dissociation.
When you said overnight wash, do you overnight wash of the system with the maintenance chip docked or with the immobilized chip docked?
As for the response level it gave the same association values but somewhat differ in the dissociation value, but not significantly.
I attached an example
Another thing I have noticed now, that even after priming the system, and when running a only a sensorgram without any injection, the baseline of the Fc2 (the immobilized channel) is decreasing in response much more the reference Fc1!!!
I mean washing with the immobilized chip. When washing overnight you wash even slow dissociating components from the chip. Starting an experiment with three blank cycles can also minimize drift in/between the channels.
Concerning your last comment. With double referencing maybe you can compensate for the differences in drift.
In the figure the baseline before injection looks pretty flat. Is the binding to the reference channel small compared to the active ligand? Since the y-axis is missing it is hard to judge the response levels.
I do not have a direct explanation for this phenomenon (yet).