6 replies to this topic

[Gradstudent78]

How much variation should one expect across a plate for measures of the same sample (using sandwich elisa)?  How much would be considered acceptable? Thanks.

[sgt4boston]

The lowest possible %CV is ideal...</=5% great </=10% good...  if you are doing the test manually you may observe an 'end of run' effect...due to the time delay from beginning to end of the assay.  Hope this helps you.

[Prep!]

10% is acceptable universally!!! some people also accept 20% It depends on ur application!!!

[Gradstudent78]

Prep!, on 26 July 2010 - 02:58 AM, said:

10% is acceptable universally!!! some people also accept 20% It depends on ur application!!!

Thanks, that's pretty much what I was thinking, 10%

[Gradstudent78]

Right now I'm getting between 10-20% across the plate when examining 10 sets of the same replicates (although the individual pairs are at CV < 10), any tips on things to do to lower this number? Thanks.

[sgt4boston]

Improving %CV:
1. automated washing
2. check calibration of pipetes
3. make sure the level of coating and reaction are different
4. with manual washing allow wash to sit 10-15 sec before decanting
5. is the number of washes sufficient
6. do you note an end of run effect.....is time critical
7. are you using a multichannel pipettor..minimize time differences

[wcw]

We use a 15% CV% cutpoint for our within-run duplicate standards.