I am validating an ELISA measuring A1c (also called glycated hemoglobin; target for humans) for use in another species. It is a direct sandwich assay. The standard and my sample (whole blood) were both serially diluted twofold.
I am wondering if anybody can help diagnose what I am seeing in my parallelism graph. As you can see in my graph (attached), the curves are near inverse of each other (linearity R2 = -0.91). Since it is a direct assay, I would of course expect this correlation to be positive.
I have gone over my plate map and how I set up my dilution rack a hundred times in my head, and I am quite certain that I did load my plate in the correct sequence as my plate map - ie, having both standard and sample increase in concentration down their respective column. Is there any other possibility (besides me incorrectly loading my plate according to my map) that could cause this "perfectly incorrect" parallelism correlation?
I would love to be proved wrong and find out that it was indeed my error in loading - but I'm pretty sure I loaded it correctly...so I just wanted to check if there was any biological answer to this (matrix effect?)
I plan to run another plate tomorrow to check- but just wanted to ask.
Thank you!
You are probably using a competitive ELISA kit. With those the signal diminishes with increasing concentration since you are measuring the labelled antibodies that are binding to sites not occupied by your target. Also be wary of using ELISA for different species as you can also get cross reactivity with proteins from the species which can interfere with the antibody binding at each step.
Robert Adolf Brinzer , Thank you for your reply. The kit says it is a direct sandwich ELISA. I ran the standard curve and samples on the same plate. How could it be competitive if the standard curve shows a direct relationship?
Sorry, thought you were testing two different kits. I would suggest trying diluting your samples further to see if the values eventually come down.
- Badges
- Science method