Maintaining Data Integrity for Western Blotting Experiments
Researchers must follow best practices for data normalization
A number of recently retracted papers related to western blot anomalies have brought the scientific publication and peer-review process under greater scrutiny, which underscores the need for data integrity stewardship and greater stringency for manuscript submission approval.
In 2023, Stanford University's former president and prominent neuroscientist Marc Tessier-Lavigne resigned from his position amid a controversy surrounding allegations of research misconduct. Tessier-Lavigne's resignation came after an eight-month investigation initiated in late November 2022, following reports by The Stanford Daily that raised concerns about potential image manipulation in his research papers dating back to 2008. Notable discrepancies included the reuse of previously published western blot images as different experiments and blot duplication across multiple experiments. Elisabeth Bik, a scientific integrity consultant and former Stanford staff scientist, has been at the forefront of raising concerns about certain papers authored by Tessier-Lavigne and others primarily revolving around potential image alterations, in particular the alteration or misrepresentation of western blot images.
The wide availability of image-manipulation tools has been of concern to the scientific community for many years now. Crowd-sourced communities have organized websites like RetractionWatch.org and PubPeer.com to flag images that have the digital signatures of image manipulation. Editors themselves are increasingly using these tools, initially developed by researchers, during the prepublication review process to ensure overall data.
However, data integrity extends far beyond protecting the publication record from fraudulent results. At its heart, data integrity is the idea that the data presented in a scientific paper accurately reflects the reality of what happened in the experiment. Experimental results can skew this reality if such experiments are not designed or executed properly. Advancing our scientific understanding depends on the accuracy of published information. It is therefore imperative for journals to have more stringent prepublication requirements, and for researchers to follow best practices for data normalization, particularly in western blotting, to prevent inaccurate data publication.
The normalization concept and a comparison of primary methods
As it pertains to western blotting, normalization can be defined as an analytical process that allows meaningful and mathematically faithful comparisons of different samples respective to a commonly shared internal control, which in turn minimizes variations. Data normalization using a loading control is required to demonstrate that a western blotting experiment was executed in an unbiased manner and that no systematic errors, such as inconsistent sample preparation, pipetting errors, and uneven gel-to-membrane transfers, have affected the experimental findings. Such experimental errors may potentially result in lane-to-lane differences in sample concentration and therefore need to be corrected before comparisons can be made among different experimental samples. Proper western blot normalization is required to show that the changes in band intensities correlate to the biological changes in test samples.
To accurately evaluate changes in target protein expression levels using a western blot, both the target protein and loading control should be measured in their linear detection range. The general criteria for a protein to serve as a loading control is that it is: ubiquitous, abundant, and constitutively expressed. Typically, “housekeeping proteins” (HKPs), such as ß-Actin, are used as internal normalization standards (i.e., loading controls) for western blots. HKPs are involved in basic cellular functions and are often highly expressed, whereas target proteins may be present in low abundance. This therefore requires a large amount of sample to be loaded to enable target protein detection. However, this also further increases the amount of HKP loaded; a discrepancy that is often not accounted for in western blot experiments, as researchers will assume that both signals are being collected inside their linear dynamic range without testing them explicitly. This practice can lead to challenges, such as the overloading of HKPs, resulting in oversaturated reference bands beyond their linear detection range. If this happens, HKPs are not serving their function as a loading control. Additionally, HKP expression levels themselves may vary under different experimental conditions and biological factors (e.g., cell cycle, cell density, tissue type, subject age, and response to treatment).
Therefore, it is important to validate any HKP for consistent expression levels for the specific sample type and experimental conditions using a Linear Range Determination test via serial dilution. This test should be performed to confirm a linear and proportionate signal response for both the HKP choice and the protein target of interest being studied. Once optimal protein loading concentration is determined, one or more rounds of optimizations may be required and normally start by adjusting the primary and secondary antibody dilution ratios.
Total protein normalization (TPN) has been introduced as a method to overcome these linearity challenges in immunodetection. TPN involves the quantification of the total amount of protein loaded in each lane, which can be measured by collecting the signal from a total protein stain on the membrane. For this method to be practical as a western blot loading control, the stain must either be compatible with immunodetection (i.e., Stain-Free technology) or reversible with destaining steps (i.e., Ponceau S). These stains tend to be less sensitive than antibody-based immunodetection, reducing the likelihood of signal oversaturation. Additionally, TPN stains exhibit good linearity within the common loading range of 10–50 μg of cell lysate per lane. This approach enables the measurement of both low-abundance target proteins (using sensitive immunodetection) and high-abundance HKPs (using less sensitive total protein staining) within their respective linear dynamic ranges. Total protein signals are also far less likely to be affected by biological changes than any individual HKP signal. TPN also eliminates the need for stripping and re-probing steps associated with HKP normalization, which reduces potential variability in the reference signal. Due to the reliable linearity and inherently robust nature of the total protein signal, there is increasing advocacy for TPN as the method of choice for western blotting normalization.
| HKP | TPN |
| Antibodies for most common HKPs are required, but readily available. | Antibodies are not required for normalization. |
| HKP expression levels are usually very high, making detection fairly easy. | HKP expression levels are inconsequential. |
| HKPs are often ubiquitously expressed. | HKP ubiquity is inconsequential. |
| Control experiments are necessary, time-consuming, and require multiple control experiments. | Control experiments are simple and only require a single control experiment. |
| HKP signal intensities are a proxy representation of all proteins present in the sample. | The total protein signal intensity is the actual signal of all proteins present in the sample. |
Stain-free total protein normalization
The advent of Stain-Free imaging technology has significantly improved TPN as an alternative loading control for quantitative western blotting.
This technology is built into Bio-Rad Stain-Free Gels and enhances the natural protein fluorescence by covalently binding to tryptophan residues in-gel via a rapid UV excitation, enabling imaging of total protein in gel and on a blot without time-consuming staining and destaining steps. It facilitates easy visualization of the experiment at every step of the electrophoresis, transfer, and incubation process, thereby allowing researchers to have high confidence that their blotting experiment is robust and reliable.
Stain-free imaging, an integral part of Stain-Free TPN, allows for the elimination of the use of problematic HKPs as loading controls, particularly in the common loading range for cell lysates. By normalizing bands to total protein in each lane, researchers can obtain more reliable and truly quantitative western blot data (Figure 1). By utilizing Stain-Free total protein measurement as the loading control, researchers can ensure that both target proteins and loading controls are assessed within their linear dynamic ranges in western blot experiments (i.e., linear response). Moreover, Stain-Free imaging eliminates the challenges associated with using HKPs as loading controls. Stain-Free total protein measurement provides a more reliable control, particularly for the loading range typically employed for cell lysates, enabling users to obtain truly quantitative western blot data (i.e., scalar response) by normalizing bands to total protein in each lane. In the case of multiplex western blots, this also means one of your fluorescent channels can be used for an additional protein of interest instead, getting more data out of one blot.
To simplify Stain-Free TPN, unique workflows have been developed for the visualization of the gel and the blot, and the imaging process does not interfere with downstream immunodetection steps.
Meeting new publication guidelines with the Stain Free Western Blotting workflow
Several prominent scientific journal publishers have recently updated their editorial guidelines for data publication, with a strong focus on reproducibility and quantitation criteria. The Journal of Biological Chemistry (JBC), for instance, provides specific submission guidelines for data collection and presentation, particularly addressing quantitative western blot publication. These guidelines offer recommendations on normalization methods and supporting data, and express concerns related to the use of HKPs, imaging technology, and antibody specificity. Adhering to best practices around data integrity has a myriad of benefits for researchers. Besides being ready to easily meet publication standards, researchers will decrease the time they spend troubleshooting failures, create new insights into how their experiments are working, and ultimately increase their confidence in their results.
| JBC Guidelines: Quantitative Blots | How to Meet Requirements |
| "Housekeeping proteins should not be used for normalization without evidence that experimental manipulations do not affect the expression." | Use TPN instead of housekeeping proteins. |
| "Methods including detection of enhanced chemiluminescence using X-ray film have a very limited dynamic range." | Use an imaging system with at least four logs of dynamic range. |
| "A description of the data supporting the specificity of all antibodies is required." | Use fully validated antibodies. |