How to Calculate Protein Concentration From A280 Using the Extinction Coefficient

By Dr. Zubair Khalid, DVM, MS, PhD · 2026-10-02

A protein solution absorbs ultraviolet light, and the amount it absorbs at 280 nm is proportional to how much protein is present. That proportionality is captured by a single number, the extinction coefficient, which converts an absorbance reading into a concentration. If you know the extinction coefficient of your protein, a spectrophotometer reading becomes a quantitative measurement in under a minute.

You will meet this calculation constantly in the lab. A NanoDrop reading before an enzyme assay, a purity check after affinity chromatography, a quick estimate before loading an SDS-PAGE gel: all of these rely on A280 and an extinction coefficient. Getting the number right matters because everything downstream (enzyme units, molar ratios, dosing in an animal study) inherits any error you make here.

Quick Answer

The extinction coefficient (ε) is the molar absorption coefficient of a protein at a given wavelength, usually 280 nm, in units of M⁻¹ cm⁻¹.
The governing relationship is the Beer-Lambert law: A = ε × b × c, where A is absorbance (unitless), b is path length in cm, and c is molar concentration in mol/L [5].
Rearranged for concentration: c = A / (ε × b). This gives molarity directly.
For mass concentration, use the Abs 0.1% value (absorbance of a 1 mg/mL solution in a 1 cm cell): c (mg/mL) = A280 / (Abs 0.1% × b).
The extinction coefficient can be predicted from sequence using ε280 = (#Trp × 5,500) + (#Tyr × 1,490) + (#cystine × 125) [1].
A rough default of ε1% = 10 (or "1 Abs = 1 mg/mL") works only as a sanity check, not as a measurement.

What the Extinction Coefficient Actually Describes

The extinction coefficient is not a fixed property of "protein" as a class. It is a property of a specific polypeptide chain, determined by which aromatic amino acids it contains and how many disulfide bonds it forms. At 280 nm, the absorbance of a protein comes almost entirely from tryptophan, tyrosine and cystine (the disulfide-bonded pair of two cysteines) [1]. Cysteine itself does not absorb appreciably above 260 nm, so reduced cysteines contribute essentially nothing at 280 nm [3].

This is why the same mass of two different proteins can give very different A280 readings. A Trp-rich protein like lysozyme absorbs strongly per milligram; a protein lacking Trp absorbs weakly. The extinction coefficient encodes that difference.

The Beer-Lambert law itself is straightforward. Absorbance is unitless because it is a logarithm of a ratio of light intensities. The extinction coefficient carries the units that make the equation balance:

$$A = \varepsilon \times b \times c$$

where A is absorbance at 280 nm, ε is the molar extinction coefficient in L mol⁻¹ cm⁻¹ (equivalently M⁻¹ cm⁻¹), b is the path length in cm, and c is the molar concentration in mol/L [5]. Solving for c gives:

$$c = \frac{A}{\varepsilon \times b}$$

Two conventions for ε coexist in the literature, and mixing them up is a common source of tenfold errors. The molar extinction coefficient (ε, M⁻¹ cm⁻¹) works with molarity. The mass-based coefficients work with mass concentration: Abs 0.1% refers to a 1 mg/mL solution, while the percent extinction coefficient ε1% refers to a 10 mg/mL (1 g/100 mL) solution and is therefore ten times larger [5]. Abs 0.1% is defined as the absorbance of a 1 g/L (0.1%, or 1 mg/mL) solution in a 1 cm cell [3]. The two are related through molecular weight:

$$\varepsilon_{\text{molar}} = \text{Abs}_{0.1\%} \times \text{MW}$$

For a protein with a known sequence, you can predict ε280 from the amino acid composition. Pace et al. (1995) derived the coefficients from 116 measured values across 80 proteins [1]:

$$\varepsilon_{280} = (n_{\text{Trp}} \times 5{,}500) + (n_{\text{Tyr}} \times 1{,}490) + (n_{\text{cystine}} \times 125)$$

The cystine term counts disulfide pairs, not individual cysteine residues. A protein with eight cysteines forming four disulfides contributes 4 × 125, not 8 × 125 [3]. ExPASy ProtParam reports two values: one assuming all Cys pairs form cystines, and one assuming all Cys remain reduced [3].

How to Calculate Concentration Step by Step

Step 1: Get the right sequence. Use the sequence of the protein that is actually in your tube. If your construct has a His-tag, GST tag or cleavage remnant, those residues are part of the molecule and affect both the molecular weight and the extinction coefficient. Entering the full UniProt precursor including the signal peptide gives you the wrong protein.

Step 2: Count Trp, Tyr and cystines. You can do this by hand from the sequence, or use a tool. The Protein Properties Calculator on this site will count them and return ε280 along with molecular weight. ExPASy ProtParam does the same and reports both the cystine and reduced-Cys values [3][4].

Step 3: Choose the right ε value. If your protein is folded and its disulfides are intact, use the cystine value. If you have reduced and denatured the protein, use the reduced-Cys value. The difference is usually small (about 1% for lysozyme) but can matter for Cys-rich proteins.

Step 4: Measure A280 with a proper blank. Blank against the exact buffer your protein is in. The NanoDrop guide suggests a buffer is suitable for A280 measurement if its spectrum varies no more than 0.04 absorbance units (10 mm equivalent) from baseline at 280 nm [5]. Buffers containing Triton X-100 or NP-9 absorb at 280 nm and are not suitable; proteins in RIPA buffer may need a colorimetric assay instead [5].

Step 5: Apply the formula. For molarity, divide A280 by (ε × b). For mass concentration, divide A280 by (Abs 0.1% × b). Both give the same answer if you use consistent units.

Step 6: Check the 260/280 ratio. A pure protein sample typically has a 260/280 ratio near 0.6, while pure nucleic acid has a ratio near 2.0 [6]. A ratio well above 0.6 suggests nucleic acid contamination, which inflates the A280 reading and therefore the calculated protein concentration.

Worked Example

Hen egg-white lysozyme (UniProt P00698) is a good test case because its sequence and extinction coefficient are well characterized. The precursor is 147 residues with a signal peptide at positions 1 to 18; the mature chain runs from residue 19 to 147 (129 residues) and contains four disulfide bonds [7].

The mature sequence is:

KVFGRCELAAAMKRHGLDNYRGYSLGNWVCAAKFESNFNTQATNRNTDGSTDYGILQINSRWWCNDGRTPGSRNLCNIPCSALLSSDITASVNCAKKIVSDGNGMNAWVAWRNRCKGTDVQAWIRGCRL

Counting from this sequence: 6 Trp, 3 Tyr and 8 Cys (which form 4 cystines) [7].

By hand, using the Pace et al. coefficients:

$$\varepsilon_{280} = (6 \times 5{,}500) + (3 \times 1{,}490) + (4 \times 125) = 33{,}000 + 4{,}470 + 500 = 37{,}970 \text{ M}^{-1}\text{cm}^{-1}$$

If all cysteines were reduced instead, the cystine term drops out and ε280 = 37,470 M⁻¹ cm⁻¹.

ExPASy ProtParam returns the same values for the mature chain: MW 14,313.14 Da, ε 37,970 (Abs 0.1% = 2.653) with cystines, and ε 37,470 (Abs 0.1% = 2.618) with reduced Cys [4]. Biopython 1.88 returns (37470, 37970) for the extinction coefficient pair and 14,313.0 Da for molecular weight.

Now suppose you read A280 = 0.75 in a 1 cm cuvette, blanked against the same buffer. Using the cystine value:

$$c = \frac{0.75}{37{,}970 \times 1} = 1.975 \times 10^{-5} \text{ M} = 19.75\ \mu\text{M}$$

Converting to mass concentration:

$$c = 19.75 \times 10^{-6} \text{ mol/L} \times 14{,}313 \text{ g/mol} = 0.2827 \text{ g/L} = 0.283 \text{ mg/mL}$$

The same answer comes from the Abs 0.1% route: 0.75 / 2.653 = 0.283 mg/mL. If the cysteines were reduced, you would get 20.02 µM = 0.286 mg/mL, about 1.3% higher.

Two things this example makes clear. First, the generic "1 Abs = 1 mg/mL" NanoDrop setting would report 0.75 mg/mL, which is 2.65 times too high for lysozyme. Second, if you had measured that same A280 = 0.75 with a 0.1 cm path length instead of 1 cm, the concentration would be 2.83 mg/mL, ten times higher. Path length is not optional information.

Reading and Interpreting the Result

The number you calculate is only as good as the assumptions behind it. A concentration derived from A280 assumes that the only thing absorbing at 280 nm is your protein of interest. Any other UV-absorbing molecule in the sample, whether a nucleic acid, a cofactor, a phenolic compound or a detergent, adds to the total absorbance and inflates your result [5]. Samples should be purified before A280 quantification.

The 260/280 ratio is your first diagnostic. For a pure protein, expect a value near 0.6 [6]. If it is substantially higher, nucleic acid is present. Light scattering from aggregates or particles also distorts the spectrum, raising apparent absorbance more at shorter wavelengths. Uncorrected spectra of nucleic-acid-free HBV capsids gave 260/280 ratios of 0.59 to 0.71, which dropped to about 0.6 after scattering correction [6]. If your ratio is off but you know the sample is clean, scattering may be the cause.

The Abs 0.1% value is often more convenient than the molar extinction coefficient because it gives mg/mL directly. Most protein ε1% values fall between 4.0 and 24.0 [5]. If you have no sequence information and need a rough estimate, ε1% = 10 is a reasonable default because a mixture of many proteins averages about 10 [5]. But this is a rough stand-in, not a measurement. For a protein with an unusual amino acid composition, the error can be large.

Reference values from the NanoDrop guide illustrate the range: BSA has ε1% = 6.67 (molar about 43,824 M⁻¹ cm⁻¹), IgG has ε1% = 13.7 (about 210,000 M⁻¹ cm⁻¹ for a 150 kDa IgG), and egg-white lysozyme has a molar ε of 36,000 to 39,000 M⁻¹ cm⁻¹ [5]. The lysozyme value you calculate from sequence, 37,970, falls right in that range.

Extinction Coefficient vs. Other Ways to Get Protein Concentration

Method	What it measures	Needs a standard curve?	Compatible with detergents?	Typical use
A280 with ε	Intrinsic UV absorbance of Trp, Tyr, cystine	No	No (many detergents absorb at 280 nm)	Purified proteins, quick checks
Bradford	Dye binding to basic and aromatic residues	Yes	Limited (some detergents interfere)	General protein quantification
BCA	Cu²⁺ reduction by peptide bonds	Yes	Yes (up to certain concentrations)	Samples with detergents
Abs 0.1% default	Assumes average protein composition	No	No	Rough estimate only

The A280 method is the only one that gives an absolute concentration without a standard curve, because the extinction coefficient is a physical property of the molecule. Colorimetric assays like Bradford and BCA depend on a reference protein (usually BSA) and assume your protein behaves like the standard, which is often not true. The trade-off is that A280 requires a purified sample and a known extinction coefficient.

The "1 Abs = 1 mg/mL" setting on a NanoDrop is not a method. It assumes that a 0.1% protein solution gives A280 = 1.0 in a 10 mm path [5]. For BSA, which has ε1% = 6.67 (Abs 0.1% = 0.667), that assumption underestimates the concentration by a factor of about 1.5. For lysozyme, it is wrong by a factor of 2.65. Use it only when you have no other information and need a ballpark figure.

Common Mistakes

Using the "1 Abs = 1 mg/mL" shortcut as if it were accurate. It assumes an average protein composition that few proteins actually have. For lysozyme, it overestimates by 2.65-fold. Use the sequence-derived ε or Abs 0.1% instead.
Forgetting the path length. A NanoDrop pedestal measurement may use a path length of 1.0 mm or less, then normalize to a 10 mm equivalent. If you are using a cuvette, confirm it is 1 cm. The same A280 with a 0.1 cm path means ten times more protein than with a 1 cm path.
Entering the full precursor sequence including the signal peptide. The signal peptide is cleaved off in the mature protein. Using the precursor changes both the molecular weight and possibly the extinction coefficient. Use the sequence of the protein that is actually in your tube, including any tags.
Counting cysteines instead of cystines. The Pace et al. equation uses the cystine term, which counts disulfide pairs. Eight cysteines forming four disulfides contribute 4 × 125, not 8 × 125 [3].
Ignoring nucleic acid contamination. A 260/280 ratio well above 0.6 suggests nucleic acid is present, which inflates the A280 reading [6]. Purify the sample or use a different quantification method.
Using a buffer that absorbs at 280 nm. Triton X-100 and NP-9 absorb at 280 nm. Proteins in RIPA buffer may not be suitable for direct A280 measurement [5]. Check the buffer spectrum against a water blank before trusting the reading.
Not reporting the ε value and assumptions. Without knowing which ε was used, whether Cys were treated as cystines or reduced, the MW, the path length and the blank, nobody can reproduce your concentration.

Limitations

The sequence-based prediction of ε280 works well for proteins that contain Trp, with an average deviation of about 3.8% overall and 3.2% for Trp-containing proteins [3]. For proteins without Trp, errors can exceed 10% [3]. Pace et al. themselves concluded that the best approach is to measure ε instead of predicting it, and recommended the Edelhoch method for measurement [1]. Gill and von Hippel (1989) reported accuracy of about ±5% for their calculation method calibrated against 18 globular proteins [2].

The method assumes no other chromophores absorb at 280 nm. Conjugated proteins with prosthetic groups, such as catalase, hemoglobin or peroxidase, cannot be quantified this way because the prosthetic group contributes to the absorbance [3].

The 260/280 thresholds for "pure" protein vary between sources. The 0.6 and 2.0 values come from Porterfield and Zlotnick (2010) and are approximate [6]. Different proteins have slightly different intrinsic ratios, and the presence of scattering particles can shift the ratio without nucleic acid contamination.

Instrument-specific numbers, such as NanoDrop path lengths and upper absorbance limits, come from one vendor guide and should be treated as examples, not universal specifications [5]. The NanoDrop guide is also internally inconsistent for lysozyme: its sample-type table lists ε1% = 24.6 while the calculation section lists 26.4. The computed value of 26.5 supports 26.4, but check your instrument's current documentation.

Biopython's molecular weight (14,313.0 Da) and ProtParam's (14,313.14 Da) are average masses that do not subtract the hydrogens lost on disulfide formation (about 8 Da for four bonds). The effect on concentration is negligible.

Frequently Asked Questions

What is the extinction coefficient of a protein?

The extinction coefficient is the molar absorption coefficient at a specific wavelength, usually 280 nm, in units of M⁻¹ cm⁻¹. It describes how much light a 1 M solution of the protein absorbs over a 1 cm path. For proteins, it depends on the content of Trp, Tyr and cystine [1].

How do I calculate protein concentration from A280?

Use c = A280 / (ε × b) for molarity, or c (mg/mL) = A280 / (Abs 0.1% × b) for mass concentration. You need the extinction coefficient or Abs 0.1% value for your specific protein, the path length in cm, and a blank reading against the same buffer [5].

What is the difference between molar extinction coefficient and Abs 0.1%?

The molar extinction coefficient (ε, M⁻¹ cm⁻¹) gives concentration in mol/L. Abs 0.1% (not the same as ε1%, which is ten times larger) is the absorbance of a 1 mg/mL solution in a 1 cm cell and gives concentration in mg/mL directly. They are related by ε = Abs 0.1% × MW [3][5].

Can I use the NanoDrop "1 Abs = 1 mg/mL" setting for my protein?

Only as a rough estimate. That setting assumes an average protein composition that few proteins match. For BSA (Abs 0.1% = 0.667), it underestimates by a factor of about 1.5; for lysozyme, it overestimates by 2.65-fold. Use a sequence-derived extinction coefficient whenever possible [5].

What 260/280 ratio should a pure protein have?

A pure protein sample typically has a 260/280 ratio near 0.6, while pure nucleic acid has a ratio near 2.0 [6]. A ratio well above 0.6 suggests nucleic acid contamination, which will inflate your A280-based concentration. Light scattering can also raise the ratio without nucleic acid being present.