There ary many reasons for this high CV.

Pipetting is one. Be aware that pipetting in ELISA is different from pipetting in other procadures like cell cultures.

1. Make sure that you didn't get any bubbles in your wells, hold the pipette tips lightslopinglyagainst the wall of the well, don't scratch the bottom of areas which could be coated the the antibody.

2. Use multichannel pipettes and calibrate your pipette channel by channel.

3. You should make sure that you are pipetting always each incubation step in the same direction with the same rhythm. If you are able to run with the same rythm, then you can run more than one plate in parallel.

4. shaking of ELISA plates (if it's not really dictated) is never helpful. You are not able to get the same conditions at each run and at each plate. Believe in the diffusion. It works, even in reality.

5. If you working with the same rhythm, do this also for your substarte reaction (Ellmann). Especially this step is very sensitive for any delays. The OD sould be always due to Lambert-Beer Law between 0.01 and 2. Remember the the absorbace is the reciprocal values of the transmission.

5. Avoid any gradients over your plate. This assays are sensitive for temperature gradients and much less for humidity (except you have long incubation times (> 12 h). To overcome this two items: Any buffer, sample, etc. should have room temperature befor you starting your assay (of if required even higher temperatures). Wells closed to the edges of the plates will get warmer much faster than the inner ones. To avoid himidity gradients, cover the plates (adhesives, an unused plate, plate covers, ...)

Dear Colleagues,

I don't know if you have some at disposer, but a way to decrease variiability ( in general due to pipetting, or lack time between first well and last, espially with 5-6 plate at the same time) is to use an automatom: either for dilutions or for distribution of reagents during the ELISA. It can be faster than human and more reproducible , if different arms act to distribute the reagents. I suppose the results depend on the kit ("robut" or not), and I don't know yours ; and also these sort of machines are not at disposure in all labs.

Regards

Didier

Dear Jamille, actually the standard curve on each plate should compensate for these variations. But try to set up a schedule where you process the plates in a fashion that each plate gets exactly the same treatment. That might help.

Anyway, still some thoughts: Do you get better results when you process one plate strictly after another, including separated sample preparation? Do you control the temperature of your reagents upon standing? Are all your plates/kits from the same lot, ordered and delivered in the same shipment, stored the same way? Do you find a correlation between the absolute ODs of your standards and plate number / incubation / waiting time?

As others have noted, staggering the times at which you start each plate, or running fewer in parallel will help. There are many other sources of variation. These include starting to use reagents before they have equillibrated to the same temperature and edge effects. See the NUNC technical bulletin on edge effects. https://static.thermoscientific.com/images/D19554~.pdf Others available on same web site.

Also, quality of plate is important. Did you observed

From KPL, Inc.: "This could be due to using substrates too soon after taking them out of the refrigerator and not allowing them to reach room temperature. Substrate turnover (i.e., the rate of color development) is temperature dependent. Therefore, a substrate should always be used at the same temperature. KPL recommends equilibrating a substrate to room temperature using a water bath (20-30 oC) or allowing the substrate to equilibrate on the bench before use."

I fully agree with the different comments but primarily I agree with the comment that you should not try to do 5 plates at once. The incubation times are critical and with that many plates you are going to have considerable variance in the incubation time.

There ary many reasons for this high CV.

Pipetting is one. Be aware that pipetting in ELISA is different from pipetting in other procadures like cell cultures.

1. Make sure that you didn't get any bubbles in your wells, hold the pipette tips lightslopinglyagainst the wall of the well, don't scratch the bottom of areas which could be coated the the antibody.

2. Use multichannel pipettes and calibrate your pipette channel by channel.

3. You should make sure that you are pipetting always each incubation step in the same direction with the same rhythm. If you are able to run with the same rythm, then you can run more than one plate in parallel.

4. shaking of ELISA plates (if it's not really dictated) is never helpful. You are not able to get the same conditions at each run and at each plate. Believe in the diffusion. It works, even in reality.

5. If you working with the same rhythm, do this also for your substarte reaction (Ellmann). Especially this step is very sensitive for any delays. The OD sould be always due to Lambert-Beer Law between 0.01 and 2. Remember the the absorbace is the reciprocal values of the transmission.

5. Avoid any gradients over your plate. This assays are sensitive for temperature gradients and much less for humidity (except you have long incubation times (> 12 h). To overcome this two items: Any buffer, sample, etc. should have room temperature befor you starting your assay (of if required even higher temperatures). Wells closed to the edges of the plates will get warmer much faster than the inner ones. To avoid himidity gradients, cover the plates (adhesives, an unused plate, plate covers, ...)

I agree with those comments for doing the right procedure and another factor also you may evaluate the ODs of your standard curve got. You may discuss with Cayman chemicals Co tech people if you have follow the manufacturer's instruction but your results are far away from the standard curve levels of manual. Because the sensitivity of EIA kits are varies in different manufacturer. Secondly, you may consider then, how the samples you collected if your standard curve levels are equal about to manufacturer's standard curve. Good luck!

THANK YOU So much everyione for all your suggestions! I followed them with the ENZO ELISA kit and on my first attempt I came out with an average of 8% interplate variability across the plates! It worked GREAT the first time! I then repeated the same process with the CAYMAN ELISA kit and got an average of 45% interplate variability!!

I have now done 4 experiments with the CAYMAN ELISA and none have given me good interplate variability…

The first one was using samples - I tested the same samples on different days = High %CV across the plates.

Second: I took a step back and did a clean experiment of just known corticosterone in buffer across 5 plates on different days = High %CV

Third: After many emails to CAYMAN tech staff I took another step back and did all 5 plates together on one day with know cort samples = High %CV

Fourth: After talking with other research techs (and all of you!) I did each plate individually one after the other and timed everything exactly the same to the MINUTE. I also moved my experiment into a cooler room and made all fresh reagents and samples. I STILL got a HIGH Interplate CV (average of 45%)!!

So needless to say I am frustrated with the amount of time and money that has gone into the CAYMAN kits.

I feel strongly that it is not my technique that is not allowing the cayman kit to work because the ENZO kit worked great for me. I feel that because the kits I ordered got stuck at the USA boarder and came to me unfrozen may be the issue. It says in the manual not to thaw and refreeze reagents and that is exactly what happened to all these 4 kits. But the CAYMAN staff assure me this is not the case as the kit was still “cool” when it came.

If it’s not the kit I am missing something substantial in this procedure. I have been trying to think of what could have gone wrong and the only things I can think that still may cause this variability are...

1) I made a large amount of each sample so I put it in a reagent boat and took each sample from there for each plate (in triplicates using a multichannel pipette) - by doing it this way I was not able to "vortex for 30 sec" before adding it to the plate. Could that have caused this huge variability?

2) My other concern is after the "microplate wash" (using a machine) there is always a tiny wet spot in the center of each well - I bang each plate out 10 times and it's still there. Should I be concerned that this will cause variability? Do I need to wait for it to dry? Normally after the 10 bangs of the plate I go immediately to adding the ellmans reagent.

3) Even though I timed everything to the minute I was not able to control the development time. I read all the plates after 1 hour and 40minutes except plate 4 because it was still not ready. So that plate took 1 hour and 55minutes. Should I be concerned that I keep taking the plates out of the dark each time to test if it's ready using the mircoplate reader?

4) Should I be concerned how the microplate washer is set for the cayman? I have it on the "NUNC-flat well setting" I wash row 1 completely with 400ul of fresh wash buffer 5 times then it goes to the row 2. The whole process takes a few minutes but maybe I should be concerned that row 1 is sitting longer in the "clean step" vs row 6 or 12??

Thanks so much everyone for any support you can give me. This is my first time using Research Gate and so far it’s been a GODSEND! :)

Jamille

HI Stephan, I am not familiar with the Cayman kit so this may be off base. I would definitely try changing the washer settings and see if that helps.

We had similar problems with another ELISA kit. After trying everything we could think of which included most of the above suggestions (which are all great) .Our problem turned out to be the 1.5 mL microcentrifuge tubes we were using during a sample pretreatment step. We had problems with the cv''s even when pipetting the same samples on the same plate. Once we changed to a different microcentrifuge tube our problems stopped.

One other thing that might be an issue is the reagent reservoirs. We have also had problems with polystrene reagent reseroirs interferring with our ELISA's in the past and now use only polypropylene.

Good Luck,

Carl

I apologize for the name confusion. I meant for this response to be for Jamille.

When I used the kit from Enzo, I had the problem that there were no signal even in TA and STD wells. Is the 500rpm 2h incubation absolutely necessary, or steady incubation could also works? Thanks in advance!