Almost every lab measures total protein before it measures anything else. Western blots are loaded by it, ELISA results are normalized to it, and enzyme activities are expressed per mg of it. The two workhorse methods, BCA and Bradford, are often treated as interchangeable. They are not. Picking the wrong one for your buffer can quietly bias every downstream number.

How each assay works

BCA (bicinchoninic acid). In alkaline conditions, peptide bonds and certain residues (cysteine, tyrosine, tryptophan) reduce Cu2+ to Cu+. Two BCA molecules chelate each Cu+ ion and form a purple complex read at 562 nm. Colour develops over time, typically 30 min at 37 °C.

Bradford (Coomassie Brilliant Blue G-250). In acidic solution the dye binds mainly to arginine, lysine and aromatic residues. Binding shifts its absorbance maximum from about 465 nm to 595 nm. The reaction is fast: 5 to 10 minutes at room temperature.

Side-by-side comparison

	BCA	Bradford
Read wavelength	562 nm	595 nm
Typical working range (standard protocol)	~20–2000 µg/mL	~125–1000 µg/mL
Micro format	~0.5–20 µg/mL	~1–25 µg/mL
Time to result	~30 min at 37 °C	~5–10 min
Detergents (SDS, Triton X-100)	Generally tolerated	Interfere strongly
Reducing agents (DTT, β-mercaptoethanol)	Interfere strongly	Generally tolerated
Chelators (EDTA)	Interfere	Mostly tolerated
Protein-to-protein variability	Lower	Higher (depends on Arg/Lys content)
Standard curve shape	Close to linear	Visibly curved

Exact tolerance limits depend on the kit, so always check the manufacturer's compatibility table for your specific buffer.

Choosing by sample type

RIPA or other detergent lysates → BCA. Bradford dye precipitates or gives a high background with SDS.
Samples containing DTT or β-ME (e.g. reducing lysis buffers) → Bradford, or a reducing-agent-compatible BCA kit.
Need a result in minutes → Bradford.
Mixed or unusual proteins (e.g. glycoproteins, arginine-poor proteins) → BCA, because Bradford response varies more between proteins.
Very dilute samples → micro BCA or micro Bradford formats, not the standard protocol.

If your buffer interferes with both, precipitate the protein (for example with acetone or TCA) and redissolve it in a compatible buffer, or dilute the sample until the interfering substance falls below the tolerated limit.

The standard curve matters more than people think

Both assays are relative. The result is only as good as the standard curve behind it.

Match the standard to the question. BSA is the default. Bovine gamma globulin (BGG) is often closer for antibody-rich samples. Report which one you used.
Prepare standards in the same buffer as the samples. This cancels much of the buffer effect.
Subtract the blank (buffer plus reagent) from every well. Do not clip slightly negative values to zero.
Do not force a straight line through a Bradford curve. It flattens at higher concentrations. A linear fit over the full range underestimates high samples and overestimates low ones. Use a quadratic or 4PL fit, or restrict the linear fit to the part of the curve that is actually linear.
Keep samples inside the standard range. Values above the top standard should be diluted and re-measured, not extrapolated.
Run standards and samples in duplicate or triplicate and check replicate %CV.

A plain-text version of the 4PL model used for curved protein standard curves:

A = D + (A0 - D) / (1 + (conc / C)^B)

A   = measured absorbance
A0  = absorbance at zero protein (lower asymptote)
D   = upper asymptote
C   = inflection point
B   = slope factor

Using protein results to normalize other assays

The most common reason to measure total protein is normalization, for example expressing an ELISA result as pg of analyte per mg of total protein.

normalized = analyte concentration / protein concentration

Two things are easy to get wrong here:

Units. Convert both values to consistent units (for example pg/mL and mg/mL) before dividing.
Error propagation. Both measurements carry uncertainty. The relative uncertainty of the ratio combines both relative uncertainties:

rel. SD (ratio) = sqrt( (SD_analyte / mean_analyte)^2 + (SD_protein / mean_protein)^2 )

Reporting only the analyte SD after normalization understates the real uncertainty.

Common mistakes

Using Bradford on RIPA lysates and wondering why the background is high.
Using BCA on samples with DTT and getting implausibly high protein values.
Reading BCA plates too early, before colour development is complete, or at different times for different plates.
Fitting a straight line to a clearly curved Bradford standard curve.
Using a different buffer for standards and samples.
Reporting protein-normalized results without stating the assay and the standard protein.

Quick decision checklist

Check the sample buffer for detergents, reducing agents and chelators.
Choose BCA for detergents, Bradford for reducing agents.
Prepare standards in the sample buffer, using an appropriate reference protein.
Subtract the blank, fit a suitable model, keep samples inside the range.
Normalize downstream results with correct units and propagated error.

Doing it in Labreador

Labreador's ELISA module includes protein normalization: enter total protein per sample and the results are expressed per unit of protein, with the uncertainty of both measurements propagated into the final value. Standard curves are fitted with 4PL or 5PL, blanks are subtracted and out-of-range samples are flagged. Everything runs in your browser, so no data leaves your machine.

If you use Labreador in a publication, please cite:

Labreador - Bioassay Analysis Platform (Version 1.6.0) [Computer software]. Zenodo. https://doi.org/10.5281/zenodo.21676846

BCA vs Bradford Protein Assay: Which One Should You Use?