How to Quantify Western Blots: A Complete Guide

Western blot quantification is one of the most common — and most commonly botched — techniques in molecular biology. Everyone does westerns. Not everyone does them quantitatively. If you've ever eyeballed a blot and said "yep, that looks like it went up," this guide is for you.

Here's how to go from a blot image to defensible, publication-ready numbers. We'll cover each step of the process, flag the mistakes that trip people up, and point you to the primary literature so you can cite your methods properly.

Step 1: Start with a Good Image

Quantification begins at the imager, not in software. The single biggest source of error is a bad image.

• Use 16-bit images. An 8-bit image has only 256 intensity levels. A 16-bit image has 65,536. That's the difference between measuring with a ruler and a micrometer. If your imager exports 8-bit by default, change the setting (Janes, 2015, Sci Signal).
• Save as TIFF, not JPEG. JPEG uses lossy compression that destroys quantitative information. TIFF and PNG are lossless. The JBC author guidelines explicitly require uncompressed images for quantification (JBC, 2024).
• Check for saturation. If any pixel in your band of interest is at the maximum value (255 for 8-bit, 65,535 for 16-bit), that band is saturated and cannot be accurately quantified. Saturated bands systematically underestimate fold changes — a real 2× difference can appear as 1.2× (Janes, 2015).
• Be cautious with film. X-ray film has a very limited linear dynamic range (often only 4–8×). Digital imaging systems are strongly preferred for any quantitative work (Janes, 2015; Butler et al., 2019, Biomed Res Int).

Step 2: Define Your Regions of Interest

You need to tell the software where your bands are. This sounds trivial, but how you draw your ROIs directly affects your numbers.

• Use consistent lane widths. Every lane ROI should be the same width. If one lane is wider, it captures more background and skews the integrated intensity.
• Band ROIs should encompass the full band. You're measuring integrated (summed) intensity over the band area, not peak intensity. A tight box around a diffuse band will miss signal (Gassmann et al., 2009, Electrophoresis).
• Include both your target and loading control. You need to measure both to normalize for unequal loading.

Step 3: Subtract Background

Every blot has background signal — chemiluminescent glow, membrane autofluorescence, or uneven illumination. You must subtract it.

The key rule: subtract background per lane, not globally. Background varies across the membrane. A single global subtraction value introduces systematic error (Janes, 2015; Taylor & Posch, 2014, Biomed Res Int).

Common methods:

• Local minimum / local median: Take the intensity from the edges of each band ROI or the median of non-band pixels in the lane. This is what most commercial software defaults to (LI-COR Image Studio, Bio-Rad Image Lab) and what Butler et al. (2019) recommend.
• Rolling ball: The Sternberg algorithm used by ImageJ. A virtual sphere rolls under the intensity surface and the contact points define background. Works well for even backgrounds, but the radius parameter matters — the default (50 px) is often wrong for your image.
• Lane profile baseline: Plot intensity vs. position along the lane, draw a line under each peak, and integrate above it. This is the method Janes (2015) recommends.

Step 4: Normalize to a Loading Control

No matter how careful you are with protein quantification and loading, there will be lane-to-lane variation. Normalization corrects for this.

The field has moved significantly in recent years. The old standard — probing for GAPDH, beta-actin, or tubulin — has major problems. Housekeeping proteins can change expression under experimental conditions (Aldridge et al., 2008, J Neurosci Methods) and saturate at surprisingly low protein loads, around 4 µg/lane (Taylor & Posch, 2014).

Total protein normalization (TPN) — using Ponceau S, stain-free gels, REVERT, or Coomassie — is now the recommended approach. The JBC, Cell Press, and Nature journals all prefer TPN over single housekeeping proteins. See our detailed post on loading control selection for the full story.

Once you have background-subtracted intensities for both target and loading control, the math is straightforward:

Normalized ratio = Target intensity / Loading control intensity

To express as fold change, divide each normalized ratio by the control sample's normalized ratio.

Step 5: Verify Linearity

This is the step most people skip, and it's arguably the most important. The ratio normalization above is only valid when signal is directly proportional to protein amount — that is, the relationship is y = mx through the origin (Butler et al., 2019).

To verify this, run a dilution series of your sample and plot signal vs. amount. If the curve bends (hyperbolic/saturating), your experimental samples must fall within the linear portion for quantification to be valid.

Step 6: Report Your Methods

Reviewers increasingly expect a detailed methods description for western blot quantification. Kroon et al. (2022, PLOS Biol) surveyed hundreds of papers and found that most fail to report basic parameters. At minimum, state:

• Software and version used
• Background subtraction method and parameters
• Normalization method (total protein, specific loading control, etc.)
• Number of biological and technical replicates
• How fold-change was calculated

Common Pitfalls to Avoid

• Quantifying saturated bands. Always check exposure before quantifying.
• Using global background subtraction. Always subtract per lane or per band.
• Assuming your loading control is constant. Validate it for your specific experimental conditions.
• Comparing across blots without a calibrator. Each blot needs an internal reference sample (Janes, 2015).
• Confusing technical and biological replicates. Three lanes from the same lysate are n=1 biologically (Kroon et al., 2022).

A Faster Way

If this seems like a lot of steps, it is. VoilaBlot automates the tedious parts — background subtraction, normalization, QC checks — while keeping you in control of the science. Your blot images never leave your browser, and you get publication-ready figures with a proper methods paragraph in minutes instead of hours.

References

• Aldridge GM et al. (2008) J Neurosci Methods 172(2):250-254.
• Butler TAJ et al. (2019) Biomed Res Int 2019:5214821.
• Gassmann M et al. (2009) Electrophoresis 30(11):1845-1855.
• Janes KA. (2015) Sci Signal 8(371):rs2.
• JBC Author Guidelines (2024). ASBMB JBC Resources.
• Kroon C et al. (2022) PLOS Biol 20(9):e3001783.
• Taylor SC, Posch A. (2014) Biomed Res Int 2014:361590.