Sandra Tolchinsky
Gal Gur

Save your samples and sanity.

A faulty antibody is not always obvious, but using one unknowingly will drain any lab’s resources, regardless of size and funding. 

Acting on the suspicion that one of their pancreatic cancer-related antibodies was faulty, a research group in Toronto launched a two-year, $500,000 investigation to discover that what they thought was an antibody against CUZD1 was actually against CA125.1

It is easy to assume that a recently purchased antibody or one borrowed from a labmate has been validated and works as intended, but there are no universal guidelines or standards for antibody production and validation. Researchers must conduct their own antibody quality control analysis to ensure their results are valid and reproducible.

Before you place an order or grab an antibody out of the freezer to start the next experiment, read through the tips below to foolproof your antibody-based experiments.

1. Seek out performance data for your intended application

Antibodies against a specific antigen do not perform equally. Even if a vendor has posted specifications and a Western blot online, it does not mean the antibody is suitable for ELISA, for example. As experimental techniques require varying antibody specificity, it is important to check that the antibody has been validated for your application before you purchase. The most reliable vendor data are those linked to published datasets on antibody performance such as the Human Protein Atlas or acquired through university screening partnerships. Reputable vendors can also provide details behind their validation practices, including the names of the positive and negative cell line controls, relevant protocols, and available journal citations.

2. Validate your antibody before first use and when you observe lot-to-lot variability

As a result of the production process, the clonality of the antibody may dictate its reproducibility.

Antibodies are produced when a host animal is challenged with the target antigen. Serum from that host contains polyclonal antibodies produced by a number of different B cells, each of which raises an antibody to recognize a different epitope on the antigen. Single B cells can also be isolated from the host’s spleen and fused to myeloma cells to create hybridomas—hybrid cells capable of secreting monoclonal antibodies specific to a single epitope.

Since they originate from a single cell that is cloned and cultured, monoclonal antibodies tend to exhibit minimal lot-to-lot variation. In contrast, each lot of polyclonal antibodies is produced by a new host immunization. Due to the nature of the mammalian immune system, each immunization will yield a slightly different arsenal of antibodies, which affects lot-to-lot reproducibility. 

One should also consider the purification method used, as some methods may leave behind impurities that could affect results for more sensitive techniques. For polyclonal antibodies, Protein A or Protein G purification yields a less homogenous product than immunogen affinity. It is best to verify in advance the level of purity required and use antibodies purified in the same manner when repeating a set of experiments.

For these reasons, it is important to validate every lot of material for your application. If a new lot fails to perform comparably to a previous lot, reach out to the vendor to see if they have made any changes to the antibody and ask for technical advice, if necessary. Discard the lot when appropriate (see #3).

The first steps of antibody validation are to perform a Western blot to verify that the antibody recognizes a protein of the expected molecular weight and then use immunocytochemistry or immunohistochemistry to verify the correct subcellular localization. Use a panel of positive and negative cell lines or tissues with variable expression levels of the protein of interest. If such lines do not exist, transfect the protein of interest in nonexpressing cells to create a positive control or use RNAi to knock down the protein of interest to generate a negative control.

In addition to validation by Western blot, the antibody must also be validated for any end application, whether it is ELISA, immunoprecipitation, or another technique.

3. Understand the meaning of multiple bands on the Western blot in the context of your experiment

Multiple bands do not necessarily mean your antibody is faulty, but they could be a cause for concern if you are not expecting them in your experiment. It is important to identify the source of multiple bands as either endogenous biology, such as multiple isoform expressions or post-translational modifications of the target protein, or as experimental artifact, such as an overloaded gel, degraded cell lysate, or a faulty antibody.

Once you have ruled out the possibility that the extra bands are scientifically relevant and related to experimental design, compare all bands with your vendor’s full Western blot image. A monoclonal antibody and a pure polyclonal antibody should ideally produce only one band for the protein of interest. The band should also be present at several antibody concentrations. If this is not the case or the reasons mentioned above are not applicable to your protein, consult your vendor and discard the antibody.

4. Be mindful of time, temperature, or storage conditions

Antibody performance degrades over time, particularly under improper storage conditions. Working directly from the original vial increases the chance of contamination and subjects the antibody to unnecessary freeze/thaw cycles. Thus, it is best to aliquot antibodies and store them on ice when working at the bench. If you are not documenting how antibodies are stored and accessed, you can never be sure of their integrity.

5. Focus on the big picture: reproducible results

Personal incentives for reproducible research are high. Publicity over a manuscript retraction—and the consequent loss of productivity—is more damaging to a researcher’s career than a vendor’s bottom line. For example, one company peddling a number of mislabeled or faulty antibodies continues to operate despite retractions, negative publicity, and investigation by the Federal Trade Commission.2

Carefully selecting your antibody and antibody vendor, as well as validating the antibody upon receipt, can save your samples and your sanity. A paper published by the Rimm Lab in Pathology department at Yale includes a useful flowchart for DIY antibody validation.3

For general antibody information, consult the following databases:

  • Antibodypedia — The Human Proteome Organization (HUPO)’s catalog of validated antibodies against human proteins (
  • Human Protein Atlas — HUPO’s high-resolution images of protein expression profiles in normal and cancerous tissue (
  • 1degreebio — Read and submit independent reviews of antibodies and antibody-related services. Vendors subscribe to have their the products and services listed on the website (
  • MIAPAR — Minimum Information about A Protein Affinity Reagent released through a working group at HUPO4 (doi:10.1038/nbt0710-650)

It is important to remember that validating reagents is only one step to ensuring reproducible results. Proper experimental design, including the use of the proper positive and negative controls, as well as orthogonal techniques to verify results, significantly improves experimental reproducibility. 

Sandra Tolchinsky ([email protected]) is a product manger of antibodies and Gal Gur ([email protected]) is a supervisor of cell biology R&D at Sigma-Aldrich. To read more recommendations from Sigma-Aldrich to improve reproducibility of experiments, visit its translational research website at

1 Prassas, I. et al. False biomarker discovery due to reactivity of a commercial ELISA for CUZD1 with cancer antigen CA125. Clin. Chem. 60, 381—388, doi:10.1373/clinchem.2013.215236 (2014).
2 Bonislawski, A. Faulty antibodies continue to enter US and European markets. Genomeweb (2013).
3 Bordeaux, J. et al. Antibody validation. BioTechniques 48, 197—209, doi:10.2144/000113382 (2010).
4 Bourbeillon, J. et al. Minimum information about a protein affinity reagent (MIAPAR). Nat. Biotechol. 28, 650—653, doi:10.1038/nbt0710-650 (2010).

Previous articleTop 20 IPOs of 2014 (So Far)
Next articleOrigami DNA Nano-Tool Sheds Light on Cancer