Lack of reproducibility using BCA to quantitate total protein - (Mar/10/2015 )

Hi. I am using a BCA assay to measure total protein in a purified sample, using the 96-well plate method from the Pierce kit. I run each BCA well in duplicate, and the duplicates are often more than 10% CV apart--drastically more. Do you know what I am doing wrong, where contamination or other issues might be occurring? Thank you for your help!

Are you following the protocol exactly as stated? How is the cv for your standards? Are you diluting your samples? Are you mixing your samples well? Could be how you dispense your sample in the well. If something was contaminated I think it would effect all standards and sample equally, meaning cv would be fine but slightly higher absorbance. Are you using a new 96well plate or one that is washed a reused? Maybe the plate reader is the problem. What is the reproducibility for blank samples?

it could also be due to timing

BCA assay is very sensitive.

Most errors in my experience are caused by poor pipetting technique and/or contamination from not wearing gloves, fingers, hair, etc.

Timing is important for BCA assays in terms of ensuring your standard curve is valid for your unknown samples, but it doesn't explain the inconsistency between two identical samples read at the same time in the same 96 well plate.

It may also be possible that your BCA stocks are contaminated from people dipping into them dozens of times with dirty technique. If the reagents are old, considering buying a new stock after ruling out errors on your end.

not just timing of the readings themselves but also timing of addition of reagents. even if using a multiple channel pipette the duplicates would have to be in the same row (or column if 12 channels) to have equal timing.

but, i agree that most differences in duplicates are caused by pipetting.

So I fixed it! The standards had low CV's, so I don't think it is because of pipetting errors. I found out that it was due to poor mixing of the sample in the BCA reagent. To fix it, I pipet up and down in the samples so they are mixed, and now the sample CV's are low and the results are reproducible. I was a little shocked that the sample showed heterogeneity, but apparently it takes some mixing to fix.