Need Help with Normalizing Western Blot Data with Inter-Gel Variability

Question

[Anto]

Hello everyone,

I am encountering an issue with normalizing my Western blot data and am seeking advice on the best normalization method to use.

Context:

I have performed several Western blots, each containing control (WT) and diseased samples. Each gel includes 1 or 2 WT controls and 2 or 3 diseased samples. The signal/loading control ratios (noise removed) vary significantly from gel to gel due to technical reasons (ranging from +1 to 0.03).

Normalization Methods Considered:

Intrawestern Normalization:

Calculate the mean ratio of the WT controls for each gel. Normalize each sample on the gel by this intragel mean.

Interwestern Normalization:

Calculate the mean ratio of the WT controls across all gels. Normalize each sample by this global mean.

Problem:

The technical variability between gels is significant, making direct comparison challenging.

The intragel method provides the most consistent results for me, as it accounts for gel-specific variability.

However, I am concerned that this method may not be rigorous enough since it does not "interact" with data from different gels.

Additionally, I cannot combine my two methods because if I take the mean of the intranormalized WT controls, the average is 1. This means that dividing the intranormalized values by this average has no effect, negating the advantage of interaction between data from different gels.

Questions:

Which normalization method would you recommend in this situation? Are there alternative approaches I should consider? Any advice on how to effectively apply a combined normalization method? Thank you very much for your help!

Answer 1

[rfran010]

Normalize the target protein to the loading controls for each sample, within each gel (normalized expression).

Using normalized expression, set your disease samples relative to the WT within each gel (relative expression).

Compare relative expressions between gels.

As far as I know, this is the only proper way.

Westerns can be very variable and that can result in different relative expression values from different gels, even with the same samples. Being consistent and strict with each western can mitigate this, but variability can always occur. If you really want to squeeze all the quantitative power out of your western, run a dilution series of the wild type samle within each gel to see how sensitive that gel is.