Hello, we have been performing in our lab a validated ELISA test and we are getting a lot of variation in results. We have a std curve with each test and we also include a control sample and we use the result of this control as a system suitability criteria to evaluate if our test is valid. The thing is that we are getting a lot of variation for the control result and we often have to reject runs because the control result is too low or too high. The tests are executed in the same conditions and with the same reagents. Maybe the room temperature can vary a little from day to day, but with a standard curve on each plate, the effect of temperature is the same on std and samples so it's not supposed to cause variation in results right? The only thing I noticed is that we seem to get lower result for the control when we full the whole plate (96 wells). So my question is: Could the time it takes to fill the 96 wells of the plate with std and sample have an effect on the final results? We begin by filling wells with 7 std concentrations in triplicate (so 24 wells including the blank) and then we fill the wells with our samples (we have to use 3 different concentrations of sample in triplicate, so 9 wells per sample). So, the time it takes to fill the wells before starting incubation is about 5 minutes longer if we have 8 samples compared to if we have 2 samples. Could it be the cause of variaton in results for the control?
Thanks for your help!
In my opinion, yes, time to fill a complete plate could make a difference. Several years back I was doing an osteopontin ELISA from scratch (coating the plate/sandwiches ourselves) and had much better results when using a multi-channel pipettor and repeater pipettor to minimize loading times. It won't help for anything where you are adding different amounts/volumes, but for wash steps and reagent steps that have each well getting the same volume, these tools are helpful.
I'd also ask (since you said "we" in the original question), do you have multiple people performing the process? I've unfortunately found that many people actually still use pipttors incorrectly - so if you have different people doing steps throughout the procedure there can be a fair amount of variation and resulting error.
Thank you for your advice! We are 2 people performing the test and we both get variation. I don't think it comes from pipetting, because we both perform successfully pipette verification and calibration. We use multichannel when adding reagents and single channel for filling wells with std and samples (that is where the filling time can differ depending if we have 1 or many samples to add). We get very similar OD for std curve each time. It is the OD for the control sample that varies from one test to another, giving us variation for final control result.
If you doing triplicates you can spread them on plate and see if there is a difference, mix well all samples. How to you wash the plate? Apply reagents.
Thanks for your advice Oded. We wash our plate with a manual device. It takes around 15 minutes to fill the whole plate with standards and samples. Yesterday, we tried to fill the plate with the samples first and then with the std (usually we fill with std first) and we saw a huge difference with the final results for samples. So, today we will try to fill the wells this way: First replicate of std, then half the samples, the second replicate of std then the rest of samples and finally the last replicate of std to see what we get.
One other item occurred to me - how are the samples prepared prior to ELISA? I'm assuming they are pre-cleared and all cell debris is spun down and the supernatant is what's run? Any chance that while the samples are sitting in the tubes, something is settling out of suspension or even precipitating?
Is there any trend to the difference in your control sample ODs? Do they become higher than you expected, lower, or are they all over the place?
- Badges
- Science topic
- Similar topics
- Biological Science
- Kinesiology
- Running