blotRig Tool Transforms Western Blot Design and Statistical Analysis for Reproducibility

By Dr. Chinta SidharthanReviewed by Laura ThomsonSep 26 2024

The reproducibility of protein analysis experiments depends on the rigor and reliability of the quantitative analysis of Western blot results. Therefore, the dearth of thorough quantitative statistical Western blot analysis methodologies poses a challenge to the reproducibility of biomolecular protein analysis experiments.

In a recent study published in Scientific Reports, scientists discussed the best practices for Western blot experiment design and analysis. They introduced the analytical tool blog, which can help design and analyze Western blot experiments with improved reproducibility and rigor.

Background

Western blotting is a technique widely used in proteomic studies to detect and measure proteins in biological samples. It is based on the interactions between antigens and antibodies. Quantitative Western blotting is conducted in four main steps: separating the proteins, transferring the separated proteins to a membrane or solid surface, labeling the proteins with antibodies, and analyzing the results.

Western blotting results are often variable due to non-linear antibody signals and differences in protein loading and gel transfer. Therefore, quantitative Western blotting is considered a semi-quantitative method because the protein amounts are relative rather than exact.

Despite advances such as fluorescent labeling techniques to improve accuracy, quantitative Western blotting continues to be variable across different labs and experiments. This variability can lead to potential errors in the interpretation of the data, impacting numerous areas of biomedical research and clinical diagnosis.

About the study

The present study aimed to improve the reproducibility of quantitative Western blots by addressing significant sources of variability, such as experimental imbalance, antibody non-linearity, and inconsistent technical replicate treatment.

Given that murine models were used in the study, the researchers followed all the ethical guidelines for the animal procedures. Non-survival spinal cord injury surgeries were performed on the rats; the rats were euthanized and the lumbar cord was extracted from them for protein lysate analysis.

The spinal cord tissue was preserved at -80 °C until the Western blot analysis was performed. The tissue samples were thawed for the protein analysis experiment, and polyacrylamide gel electrophoresis was used to separate the proteins. The study focused on quantifying the levels of the alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor subunit glutamate ionotropic receptor 2 (GluA2), with beta-actin as the loading control.

The experimental process for the Western blot analysis involved a series of carefully controlled steps involving randomization, counterbalancing, and numerous technical replicates to minimize variability. The entire process was also repeated three times. The bicinchoninic acid (BCA) assay was conducted for protein quantification.

Multiplexed near-infrared immunoblotting was used for the Western blot analysis. The protein samples were diluted and loaded onto a nitrocellulose membrane and incubated with primary antibodies targeting GluA2, GluA1, and postsynaptic density protein 95 (PSD95). Subsequently, fluorescent-labeled secondary antibodies detected the protein-antibody complex and were imaged.

Near-infrared densitometric analysis was also employed to quantify the protein levels. Compared to traditional chemiluminescence methods, this method was considered more sensitive and specific, allowing multiple proteins to be detected simultaneously. The researchers also introduced the tool blotRig to improve the balancing of samples and statistical analyses for enhanced reproducibility.

Major findings

The study found that the intensity of the Western blot band and the protein concentration have a non-linear relationship, particularly at extreme values. If linear models are used, this can lead to statistical errors and inaccurate protein quantification. Therefore, it is important to determine the linear range for each antibody to ensure that the band density changes reflect the protein concentrations.

The counterbalancing process, where the experimental groups are evenly distributed across gels, was essential for avoiding confounding the technical variability from the gels with biological differences between experimental groups. The researchers recommend running triplicates of all samples across separate gels to reduce technical variability and increase sensitivity.

The researchers also recommended changes in the statistical approaches to improve the power and accuracy of Western blot experiments. This included treating the replicates as random effects in linear mixed models and loading the control as a covariate to account for the technical variability between gels and allow the biological effects to be interpreted accurately.

The study introduced the blotRig software, simplifying Western blot experiments' design, replication, and analysis. It included detailed information with examples of the blotRig Gel Creator interface and a comprehensive workflow for the statistical analysis of replicate data from Western blot experiments.

The software allows users to upload CSV files containing experimental groups and subject identification and balances the samples across gels to reduce variability. It ensures that each gel has a representative sample from each group, and the layout is designed so that samples from the same group are not next to one another. The users can then input their western blot data and blotRig analyzes it using linear mixed models, providing the results as p-values, graphs, and confidence intervals.

Conclusions

Overall, the study emphasized the importance of rigorous design and statistical analyses to improve the accuracy and reproducibility of Western blot experiments. The authors recommended balancing sample sizes across gels, conducting the experiments in triplicates, and using linear mixed models to make the results reliable and reproducible.