How to Quantify Western Blot Bands

1

What is densitometry?

Densitometry measures the optical density of a band on your blot — essentially, how dark it is. Darker bands contain more protein. The software sums the pixel intensity values within a defined band region to give you a single number that represents the amount of protein in that lane.

This number on its own is not meaningful. It depends on exposure time, membrane quality, antibody concentration, and how much protein you loaded. That's why every quantification workflow involves two essential steps: background subtraction and normalisation.

Key principle

Densitometry is always relative, never absolute. You're comparing band intensities between conditions, not measuring actual protein mass.

2

Image requirements before you start

The quality of your quantification is only as good as your image. Before you start measuring, check the following:

Requirement	Why it matters
Not overexposed	Saturated pixels all read as the same maximum value, so you can't distinguish differences in band intensity. If any part of the band is pure white (on a dark background) or pure black (on a light background), the image is overexposed and cannot be reliably quantified.
Unedited original	Always quantify from the raw image file, not a JPEG exported from a figure or an image that has been contrast-adjusted. Processing changes pixel values.
Consistent format	Use TIFF or PNG where possible. JPEG compression introduces artefacts that affect pixel values.
All lanes on one image	Lanes compared to each other must come from the same membrane and exposure. Splicing lanes from different blots together and then quantifying is not valid.

Watch out

Many western blot imaging systems apply auto-contrast when displaying or exporting images. Make sure you are exporting the raw data file, not a display-adjusted version.

3

Measure band intensity

Draw a region of interest (ROI) around each band. The software sums all the pixel values within this region. This raw sum is called the integrated density or raw band intensity.

ROIs should be the same size for every band you're comparing. If you draw a larger box around one band, you capture more background pixels and inflate that band's reading relative to others.

Raw integrated density Raw intensity = Σ (pixel value) for all pixels in the ROI

Tip

In AutoBlot Studio, band detection is automated — ROIs are placed consistently using the same algorithm for every lane, removing the human bias that comes from manually drawing boxes.

4

Background subtraction

Even in an empty region of your membrane, there will be some non-zero pixel values — from non-specific antibody binding, membrane fluorescence, or noise. This background needs to be subtracted from each band measurement.

The most common approach is the rolling ball or local background method: measure the pixel intensity in a region immediately adjacent to each band (above and below, or to the side), average it, and subtract it from the band ROI value.

Background-corrected intensity Corrected intensity = Raw intensity − (mean background pixel value × ROI area)

Common mistake

Using a single background region from one corner of the blot for every band. Background is not uniform across a membrane — measure it locally, adjacent to each band.

5

Normalise to a loading control

Even with careful pipetting, you will never load exactly the same amount of protein in every lane. Normalisation corrects for this by dividing your target protein intensity by a loading control measured in the same lane.

There are two main approaches:

Total protein normalisation (recommended)
Stain the membrane for total protein (e.g. Ponceau S, REVERT, or a fluorescent total protein stain) and use the total protein signal in each lane as the denominator. This is now considered the gold standard by most journals.

Total protein normalisation Normalised value = Corrected target intensity ÷ Total protein intensity (same lane)

Housekeeping protein normalisation
Use a separate antibody to detect a constitutively expressed protein such as β-actin, GAPDH, or α-tubulin. This is widely used but has important limitations — housekeeping protein expression can vary with treatment, cell type, and passage number.

Housekeeping normalisation Normalised value = Corrected target intensity ÷ Corrected housekeeping intensity (same lane)

GAPDH and β-actin caveats

GAPDH expression changes under hypoxia, serum starvation, and many metabolic perturbations. β-actin varies with cytoskeletal treatments. If your experiment involves any of these, validate your housekeeping protein or switch to total protein normalisation.

6

Calculate relative expression

Once you have normalised values for each lane, express them relative to a reference condition — typically your untreated control or a vehicle lane. Set the control to 1 (or 100%) and express everything else as a fold-change relative to it.

Relative expression (fold change) Fold change = Normalised value (sample) ÷ Mean normalised value (control group)

Repeat this for each biological replicate independently — meaning each replicate gets its own fold-change value calculated from its own control lane on the same blot. Do not pool raw intensities across blots and then calculate fold-change once at the end.

Tip — n=3 means 3 blots

A biological replicate is an independent experiment, not three lanes on the same blot. Three lanes on one blot is a technical replicate and gives you a measure of pipetting precision, not biological variability.

7

Which statistical test to use

Once you have a fold-change value from each biological replicate, you have a dataset you can apply statistics to. The right test depends on your experimental design:

Design	Recommended test	Notes
2 groups (e.g. control vs. treated)	Unpaired t-test	Assumes normal distribution. With n≥3 this is generally acceptable for western blot data.
2 groups, paired samples (e.g. before/after in same animal)	Paired t-test	More powerful when measurements are matched. Use when each control has a corresponding treated sample from the same experiment.
3 or more groups	One-way ANOVA + post-hoc test	Follow with Tukey's (all pairs), Dunnett's (all vs. control), or Bonferroni correction. Never do multiple t-tests without correction.
2 variables (e.g. treatment × time point)	Two-way ANOVA	Tests main effects and interaction. Use Tukey's or Sidak's for post-hoc comparisons.
Small n or non-normal distribution	Mann-Whitney U (2 groups) or Kruskal-Wallis (3+ groups)	Non-parametric alternatives. Use when you cannot assume normality, especially with n<5.

What p-value threshold to use

The conventional threshold is p<0.05, but report the exact p-value, not just "significant" or "ns". Reviewers increasingly expect this. Also report effect size (fold change) alongside the p-value — a statistically significant result with a tiny effect size may not be biologically meaningful.

Western blot data is often not normally distributed, especially with small n. If you have fewer than 5 biological replicates, consider reporting individual data points rather than mean ± SEM, and use a non-parametric test.

What to show in your figure

Plot individual data points alongside the bar or show a dot plot. Mean ± SEM with n=3 hides whether your replicates are consistent or wildly variable. Most journals now require or strongly prefer showing individual data points.

8

Common mistakes and how to avoid them

Mistake	How to avoid it
Quantifying an overexposed image	Check for saturated pixels before measuring. Take multiple exposures and use the one where bands are clearly visible but not saturated.
Using JPEG images	Always export as TIFF or PNG from your imaging system. Never quantify a JPEG that has been re-saved.
Different ROI sizes per band	Use consistent, identical ROIs. AutoBlot Studio automates this.
Single background measurement for the whole blot	Measure background locally, adjacent to each individual band.
Treating technical replicates as biological replicates	Each blot run on a different day from independently prepared samples = one biological replicate.
Running multiple t-tests without correction	Use ANOVA with a post-hoc correction when comparing 3 or more groups.
Normalising to a housekeeping protein that changes	Validate your housekeeping protein under your experimental conditions, or switch to total protein normalisation.

What is densitometry?

Image requirements before you start

Measure band intensity

Background subtraction

Normalise to a loading control

Calculate relative expression

Which statistical test to use

Common mistakes and how to avoid them

Try it yourself in AutoBlot Studio