Western blot quantification: from blot image to statistics
Western blot quantification turns band images into numbers you can compare: measure each band’s background-subtracted volume, normalise it to a loading control in the same lane, express it as fold change over control, and test on biological replicates. This guide walks through each step and shows how to do it in the free LabTools Gel Analyzer.
In this guide
- 1. Capture an image you can quantify
- 2. Find the lanes and measure the bands
- 3. Normalise to a loading control
- 4. Replicates and statistics
- 5. Plot and report
- Frequently asked questions
Normalised band intensity for four conditions, all 12 lanes shown, with tests on biological-replicate means (n = 3). See the full example
1. Capture an image you can quantify
Quantification is only as good as the image. Use the original file from the imager, ideally a 16-bit TIFF, not a JPEG, a screenshot or a figure that has been contrast-adjusted. Choose an exposure where the strongest band you want to measure is not saturated: a clipped band reads too low and its volume cannot be used.
In LabTools, open the Gel Analyzer and drop the image in (TIFF, PNG, JPG or BMP), or use Open from LabTools for an image already in your files. The Image step reports the bit depth, how much of the image is saturated and whether bands are dark on light (most Westerns) or light on dark. Crop and straighten here if needed.
2. Find the lanes and measure the bands
On the Lanes step the lanes are found for you. Fix the count in How many lanes? if needed, drag lanes on the image to move or tilt them, name each lane and set its role: sample, ladder, loading control or blank.
On the Bands step every band is measured. The lane profile shows the signal from top to bottom; the dashed line is the background taken off and the shaded areas are each band’s volume. If the baseline cuts into bands, try another Background method (rolling ball plus lane baseline is the recommended default). Click in a lane to add a band, drag its edges to change what is counted; edited bands turn yellow.
With a ladder lane, the Ladder step sizes every band (bp or kDa) from built-in DNA and protein ladders or your own sizes, and shows the fit with its R².
3. Normalise to a loading control
Raw band volume depends on how much protein went into the lane. On the Results step, Quantify (Western blot) takes the target band and a loading control (one click picks the same band in every lane) and divides one by the other:
- Normalize by → Loading control (same lane): a housekeeping band such as GAPDH or actin.
- Normalize by → Total lane signal: the whole lane, a total-protein style normalisation.
- Fold change against: the control lane, so every value reads as a multiple of control.
To confirm you are in the linear range, use Check the linear range with a dilution series: type the relative amount loaded in each dilution lane and LabTools reports how straight the signal-versus-load line is for the target and the loading control.
4. Replicates and statistics
The unit of analysis is the biological replicate, not the lane. Lanes from the same lysate are technical repeats: average them first, then compare conditions on the replicate means. Treating 12 lanes as n = 12 when there are 3 biological replicates is pseudoreplication and makes p-values far too small.
The Gel Analyzer’s Replicates button combines several saved gels (one gel per replicate), averages lanes with the same label on a gel, and gives the mean, SD and n per condition with a Welch t-test of each condition against the control, Holm-adjusted. Batch applies one saved analysis to a set of images so every gel is measured the same way.
With three replicates per group, rank tests cannot reach significance (a Mann–Whitney test on 3 vs 3 cannot go below p = 0.081), so a Welch t-test with a correction for the number of comparisons is the usual choice. See the western blot example for the full set of tests on real-shaped data.
5. Plot and report
Show every data point, not only a bar: a box plot or a bar with points makes the replicate structure visible. Send the band table to Charts & Plots, choose Box Plot, and use Statistical Testing to add brackets with Welch’s t-test and a Holm-Bonferroni or Benjamini-Hochberg correction. Export at 300–600 dpi or as SVG/PDF.
In the methods, state the image source and bit depth, the background method, the normalisation, the number of biological and technical replicates and the test. The example page has a methods paragraph you can adapt.
Frequently asked questions
How do you quantify a western blot?
Measure each band’s volume (its signal summed over the band, with the background taken off), divide it by a loading control from the same lane, express it relative to the control condition, and compare conditions on biological replicates. Use the original image from the imager and make sure no band you quantify is saturated.
Housekeeping protein or total protein for normalisation?
Both are used. A housekeeping band (GAPDH, actin, tubulin) is simple but can saturate or change with treatment; total protein in the lane is often more robust. Whichever you use, it has to be in its linear range too. LabTools offers both as “Loading control (same lane)” and “Total lane signal”.
How many replicates do I need for a western blot?
At least three biological replicates (independent samples or experiments), each on its own gel or lane set. Technical lanes from the same lysate are averaged first. With n = 3 per group only large effects are detectable: in our example the minimum detectable effect at 80% power is d = 3.07.
Can I quantify a JPEG or a screenshot of my blot?
Avoid it. JPEG compression and screenshots change pixel values and often clip the bright end, so volumes become unreliable. Export the original file from the imager, ideally a 16-bit TIFF; the Gel Analyzer reads it at full bit depth.
What is the linear range and how do I check it?
It is the range of protein loads over which band signal rises in proportion to the amount loaded. Load a dilution series of one sample (for example 1, 2, 4, 8 parts) and check that signal against load is a straight line. The Gel Analyzer has “Check the linear range with a dilution series”, which reports R² (pass mark 0.98 by default).