High variation among triplicates in ELISA

[pilpel]

Hey all,
So I have decided to do an ELISA.

I am using IFNgamma as the capture Ag, anti IFNgamma as the antibody and a 1:6000 diluted Protein-A HRP which serves as the detecting factor. ON top of that, I overlay with a KPL ECL mix (which I know to work excellently in other assays I've done with it).
What I've seen is usually a variation of 1 out of 3 triplicates, though not all wells suffer from this huge variation: 21253, 22494, 13908 (as an example of a 30% variation).
Now, I can do a quadruplicates and even pentaplicates and take out the best 3, but I try to refrain from doing so. Does anyone has a clue as to what be the cause to this problem?

Thanks,
CHG

[biohazard]

I assume that you have got your pipetting and washing techniques correct, and that your other ELISAs work just fine. And if that is the case, it indeed sounds pretty odd that only this one kit has so much variation. I'm not sure how often this has happened (e.g. a few failed assays can be due to just plain poor luck and you've made an error in pipetting or with washing etc.), but if the same keeps happening again and again then probably it's something with the kit, equipment or solutions used.

Maybe you should try out some other manufacturer's kit or alternative IFN-g protocols that e.g. have different blocking protocols. I know poor replicates are difficult to explain with faults in the kit or protocols themselves, since then the variation would probably take place between different ELISA runs, rather than within the same run.

It could be something small in your technique and the IFN-g kit just happens to be more sensitive to it (e.g. problems with washing), so I'd try to alter the protocol a bit, change the kit or ask someone else to do the ELISA for you, and see if they get it right. If they do, look what they did differently from your own way to do it