[raney nickel]

Ok, I have some standard curves to analyze by two methods BCA assay and Bradford Assay.  I have to compare the two results to find an original concentration of an unknown, then test the reliability of this answer.  I used standard curve line drop to estimate concentrations of my dilution scheme samples for both experiments.  I got two different sets of answers for the concentrations for BCA and Bradford. They're close relatively, but each give different answers for the concentration. 

So my question is this, how do I compare these two sets of data to find the original concentration of unknown?  And how can I be confident with one result?

[Babcock_Hall]

The Bradford assay does not detect all proteins with exactly the same response; there is a dependence on the amino acid composition.  I am not sure what affects the relative response of the BCA assay.  However, I am not surprised that the two values do not exactly agree.

[raney nickel]

Thank you for the response.  Yes, you are right about the Bradford.  Coomassie dye binds to acidic and basic residues.  With the BCA, tyr and trp residues reduce Cu+2 --> Cu+1 , which forms a complex to BCA and shifts the absorbance to 562. 

I guess what I'm trying to ask for is which value do I suggest is the more accurate for my unknown?  I want to kind of compare these two sets and find which one produced better results.  Since they're different sets of data, is there some kind of statistical test I can employ that can compare these apples and oranges?  Or is that even the right way to think of it? 

[Babcock_Hall]

I am not sure that you are entirely correct about the chemistry of the BCA assay.  I am given to understand that the peptide bond provides the electrons to reduce copper, although the amino acids you mention do have some influence.  "The reaction that leads to BCA color formation as a result of the reduction of Cu2+ is especially influenced by the presence of three particular amino acid residues in proteins: cysteine/cystine, tyrosine and tryptophan. Apparently these amino acids enhance copper reduction independently and in the biuret reaction, thereby causing formation of a colored BCA-Cu1+ chelate. However, studies performed with di- and tripeptides indicate that these produce more color than can be accounted for by the four individual BCA-reactive amino acids. In other words, the peptide backbone (and thus the total amount of protein) is the major contributor to the reduction of copper in the biuret reaction and color development in the BCA assay. Slight protein-to-protein variation in the BCA protein assay results from differences among proteins in composition with respect to these three amino acids."
http://www.piercenet.com/method/chemistry-protein-assays

I am guessing that this is for a laboratory assignment of some kind.  Did your instructor provide any further guidance?

[WW]

Just use the same method for the standard curve that you're going to analyze your unknown with

[raney nickel]

The question is more one of testing the reliability and tolerance of both methods.  I'm not having an issue with the standard curves per se, but rather choosing one method over the other.

Both methods (BCA and Bradford) were used to test the same protein, but yielded slightly different results.  The results are similar and in the ball park of each other (with similar standard deviations), but could potentially yield significant error for more precise work.  I want to tackle the questions of accuracy and precision with this query. 

Is there a way to test both sets of data to determine which was the better, more accurate result? 

[Babcock_Hall]

Make sure that you understand the difference between accuracy and precision.  Also, a great deal depends upon whether you make a standard curve using the unknown protein or a commercially available protein.  The latter is more common in my experience.  See p. 10 of this pdf file:  http://wolfson.huji.ac.il/purification/PDF/Protein_Quantification/BIORAD_Protein_Assay.pdf