Skip to main content
PMC home page
F1000Research logoLink to F1000Research
. 2013 Jul 10;2:153. [Version 1] doi: 10.12688/f1000research.2-153.v1

Reporting research antibody use: how to increase experimental reproducibility

Matthew A Helsby 1,2, Joe R Fenn 1,2, Andrew D Chalmers 1,2,a
PMCID: PMC3829129  PMID: 24358895

Abstract

Research antibodies are used in a wide range of bioscience disciplines, yet it is common to hear dissatisfaction amongst researchers with respect to their quality. Although blame is often attributed to the manufacturers, scientists are not doing all they can to help themselves. One example of this is in the reporting of research antibody use. Publications routinely lack key details, including the host species, code number and even the company who supplied the antibody. Authors also fail to demonstrate that validation of the antibodies has taken place. These omissions make it harder for reviewers to establish the likely reliability of the results and for researchers to reproduce the experiments. The scale of this problem, combined with high profile concerns about experimental reproducibility, has caused the Nature Publishing Group to include a section on antibody information in their recent Reporting Checklist for Life Science Articles. In this commentary we consider the issue of reporting research antibody use and ask what details authors should be including in their publications to improve experimental reproducibility.

Antibody information is routinely omitted from publications

Neuroscience, cancer research, regenerative medicine, infection and immunity, cell biology and cardiovascular research are just some of the fields in which research antibodies are commonly used. The sheer scale of their use is illustrated by huge sales, estimated to be worth in excess of $1.6 billion annually 1. Despite, or perhaps because of, this widespread use, it is common to hear dissatisfaction among research scientists about the quality of these antibodies. The finger of blame is often pointed at the manufacturers 2, yet it is questionable whether scientists themselves are doing everything they can to help the situation; surely not all problems can be placed at the door of the antibody manufacturer. One example of scientists not helping themselves is in their reporting of antibody use. There are many cases of good practice and detailed reporting, but all too frequently authors omit key details. These include the host species and code numbers, but even the source of the antibody may be left out. This makes it harder for reviewers to establish how well characterised the antibodies are and thus how reliable the data presented are likely to be. It also makes it more difficult for other researchers to accurately reproduce experiments.

Failure to report key information is not a new problem 2, 3, but recent developments have increased efforts to find a solution. In particular, experimental reproducibility has been thrust into the limelight by high profile cases. For example, a study of "landmark" cancer research papers found that scientific findings from only 11% of them could be repeated 4. Taken at face value this is a shocking statistic and, in an attempt to try to improve experimental reproducibility, the Nature Publishing Group have recently introduced a reporting checklist for life science articles 5. This checklist highlights research antibodies as a reagent type for which reporting could be improved. A key question is: what information to provide? In this commentary we consider what information authors should be including in their publications to help improve experimental reproducibility.

Key details for reporting antibody experiments

Publications need to report core information regarding the antibodies that were used. This should include the name of the antibody, the company/academic who produced the antibody, the host species in which the antibody was raised and whether the antibody is monoclonal or polyclonal. In addition, the catalogue or clone number needs to be mentioned. This information is commonly omitted from current publications, but is important as large antibody companies will often have multiple antibodies to the same target; a unique identifier is therefore essential to allow unambiguous identification of the antibody concerned. For this reason the first step in improving reporting should be to make it mandatory for authors to include core antibody information, including a code or clone number for the antibodies they use.

A second type of information that should be reported relates to experimental details. The application the antibody was used for is of central importance. This information is normally present, but it can be hard to extract if the antibody information is listed in a ‘Materials’ section and separated from descriptions of the techniques. Having the antibody data and application data closely linked would avoid potential confusion. Furthermore, if a study uses samples from more than one species then it is also important to clearly link which antibodies were used in which species.

There are other features that could also be reported which may be particularly relevant to certain studies. For example, the antibody batch number is rarely reported, but there is evidence of variability between different antibody batches 6, 7. This type of variability is likely to be a particular issue with polyclonals 2, but may affect monoclonals 8. Reporting the final antibody concentration or dilution is another piece of information which can help other researchers, especially if optimisation was required during the study. Finally, it has been proposed that scientists should know the antigen which was used to raise the antibody 3. This information may be commercially sensitive, but at least the location of the antigen within the protein should be known, as it will have implications for interpreting the results of certain studies. In these cases authors should be encouraged to report antigen location.

Antibody validation

The Nature Publishing Group checklist requires authors to demonstrate that every antibody used in their study has been validated for use in each of the species and specific experiments used. Validating an antibody is a complex process worthy of its own review 9 and reporting it can be achieved in a number of ways. Supplementary information could be included to demonstrate validation by the author or a citation could be given to highlight a previous study in which the antibody was validated. Reference to the antibody validation profile from publically available databases such as 1degreebio, Antibodypedia, CiteAb or pAbmAbs could also be used. Including this information would help reviewers and other researchers accurately assess the results.

A simple format for reporting antibody information

Based on the points discussed above we would suggest researchers use the following format for reporting antibody information:

"The following antibodies were used, Mouse anti-protein A monoclonal antibody (company E, catalogue number #1000) was used for Western blotting with human cells, as validated in (figure X or reference Y or validation profile Z) and Western blotting in mouse tissue as validated in (figure X or reference Y or validation profile Z). Goat anti-protein B polyclonal antibody (company F, catalogue number #1001) was used for ELISA in human tissue as validated in (figure X or reference Y or validation profile Z) and flow cytometry in human tissue as validated in (figure X or reference Y or validation profile Z)".

This format is meant as a guide and could be adapted as required; for example, details of batch number, dilution or epitope could be added where particularly important. This information could also be usefully presented in a table if allowed by the journal. Adoption of these reporting guidelines will not eliminate researchers’ frustrations with antibodies, but should help improve experimental reproducibility and scientists’ productivity, something we all seek. An additional benefit for authors who include this information is that well annotated publications are easier for antibody companies and antibody search engines like CiteAb to highlight in their databases. This inclusion is likely to increase the number of researchers who access their work and so potentially the impact of the study.

A final thought is that journals have a big role to play in promoting good practice by including guidelines on reporting antibody details in their instructions to authors and encouraging reviewers to consider this aspect of publications when they carry out their review.

Funding Statement

ADC and MH are funded by a Higher Education Innovation Fund grant (#HIF36 Chalmers) from the University of Bath, Research Development and Support Office.

The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

v1; ref status: indexed

References

  • 1.Bird C: Antibody User Survey.The Scientist,2012. Reference Source [Google Scholar]
  • 2.Couchman JR: Commercial antibodies: the good, bad, and really ugly. J Histochem Cytochem. 2009;57(1):7–8 10.1369/jhc.2008.952820 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Saper CB: An open letter to our readers on the use of antibodies. J Comp Neurol. 2005;493(4):477–8 10.1002/cne.20839 [DOI] [PubMed] [Google Scholar]
  • 4.Begley CG, Ellis LM: Drug development: Raise standards for preclinical cancer research. Nature. 2012;483(7391):531–3 10.1038/483531a [DOI] [PubMed] [Google Scholar]
  • 5.Reducing our irreproducibility. Nature. 2013;496(7446):398 10.1038/496398a [DOI] [Google Scholar]
  • 6.Chalmers AD, Pambos M, Mason J, et al. : aPKC, Crumbs3 and Lgl2 control apicobasal polarity in early vertebrate development. Development. 2005;132(5):977–86 10.1242/dev.01645 [DOI] [PubMed] [Google Scholar]
  • 7.Pozner-Moulis S, Cregger M, Camp RL, et al. : Antibody validation by quantitative analysis of protein expression using expression of Met in breast cancer as a model. Lab Invest. 2007;87(3):251–60 10.1038/labinvest.3700515 [DOI] [PubMed] [Google Scholar]
  • 8.Voskuil J: The troubles with commercial research antibodies dissected. Everest Biotech Blog. 2013. Reference Source [Google Scholar]
  • 9.Bordeaux J, Welsh A, Agarwal S, et al. : Antibody validation. Biotechniques. 2010;48(3):197–209 10.2144/000113382 [DOI] [PMC free article] [PubMed] [Google Scholar]
F1000Res. 2013 Aug 1.

Referee response for version 1

Simon Glerup 1

This commentary is much needed in the field of life science. It is well written and concise. Andrew Chalmer’s group has contributed significantly to the use of research antibodies by creating CiteAb. When operating the CiteAb search engine, I imagine that they constantly run into problems with publications with poorly described use of research antibodies.

I have two minor suggestions:

  1. In the Antibody Validation paragraph, a statement could be included in the methods section of a paper regarding if, where and under what name antibody validation information or reviews has been posted in publically available databases. This would increase the value and transparency of these databases.

  1. Unlike the previous reviewer, I think it is fine to mention CiteAb in the paper. After all, even Nature Publishing Group is a highly commercial enterprise. However, I suggest that a table could be included listing the relevant databases including CiteAb, pAbmAbs, Biobrea, Antibodypedia, 1degreebio, Antibody-Advizer etc. In this regard, I regret that the Checklist from Nature Publishing Group only refers to sites in which they have a commercial interest (1degreebio and Antibodypedia). I hope that other publishing groups are not tempted to do the same.

I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.

F1000Res. 2013 Sep 3.
Andrew Chalmers 1

We thank Professor Glerup for his helpful comments and share his concern about the Nature Publishing Group guidelines. We explain our response to each one in turn below.

  1. The fact that if no previous validation has occurred then it should be carried out and reported and/or submitted to a public database has been made clearer.  

  1. This is a good point and we agree it is important to give an overview of available databases to allow readers to choose the most appropriate. For this reason we were careful to mention a range in our first version. However, we feel it would not be appropriate for us to compile a table given our clear affiliation to one database, instead we provide a link to the most complete list of databases we are aware of.

F1000Res. 2013 Jul 24.

Referee response for version 1

John Colyer 1

The title and abstract are clear and appropriate. The article is timely and written clearly and accessibly.

  • It could be improved further by providing references for papers that are examples “of good practice and detailed reporting”, which might serve as a template for others.

  • The process of antibody validation is worthy of more extensive discussion, as the research community needs to develop a clear understanding of the most appropriate tests to be performed in each experimental system, and standards which should be attained for acceptance of the status of “validated”.  This data should be provided in supplementary data, or by reference to previous supplementary data if the same reagents are used in a new study.

  • The importance of batch number is made, but could be emphasized more.

  • Finally, the critical role of peer-reviewers in evaluating and enforcing these standards is key. Some discussion of this would enhance the manuscript.

I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.

F1000Res. 2013 Sep 3.
Andrew Chalmers 1

We thank Professor Colyer for his positive and helpful comments and explain our response to each one in turn below.

  • A good idea, we have now added an example reference that illustrates good reporting practice (Antibody information is routinely omitted from publications’ section). Journals which already encourage good practice have also been highlighted (‘Change will require help from journals and reviewers’).  

  • We completely agree and have increased the amount we cover on this topic, but not attempted a full review as we feel such a complex topic is beyond the scope of this comment article. We have added some additional citations for readers who require more information (Antibody Validation section).

  • The fact that if no previous validation has occurred then it should be carried out and reported and/or submitted to a public database has been made clearer. The fact that previous validation can be cited has also been spelled out (Antibody Validation section).

  • Additional emphasis has been added regarding the problem of batch to batch variability (‘Key details for reporting antibody experiments’ section).

  • This has been added to the ‘Change will require help from journals and reviewers’ section.

F1000Res. 2013 Jul 16.

Referee response for version 1

David Soll 1

This commentary is timely and well written, but it could be shortened or tightened up a bit for the purpose of conciseness.  It also should include a few points noted in this review.  The major point is the problem that lack of information in publications involving research antibodies affects assessment and future use.  The discussion could be more efficient in stating that if methods were reported in a previous referenced article, then referencing that article in a new publication is sufficient, unless there are nuances (i.e., new uses of the antibody).  It should also be made clear that such information be mandatory when an antibody is used in a particular way for the first time. 

There are also a few things the author may want to include:

  1. Many antibodies work on a particular protein in a particular cell type without knowledge of the protein domain(s) found.  In spite of that they may be of value, so you don’t have to identify the sequence molecule. 

  1. Some antibodies identify native conformation and therefore are not on a peptide sequence per se.  Such antibodies are not unusually performed on denatured proteins in western blots, but may work in nature gels. 

  1. Some antibodies have not been fully characterized beyond reference to the data sheet provided by the company or source if necessary. 

  1.  If the authors of a paper refer to the company, and catalog name of the antibody, prior characterization can access.

  1. Antibody validation should go in the supplementary data to a paper.

  1. Do not cite CiteAb in your paper - it sounds like an ad. 

But all in all, this is a reasonable commentary.  It reinforces what many already are advocating. The title and abstract were fine.

I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.

F1000Res. 2013 Sep 3.
Andrew Chalmers 1

We thank Professor Soll for his positive review and helpful comments. We have now addressed them and explain our response to each one in turn below:

‘The discussion could be more efficient in stating that if methods were reported in a previous referenced article....’

The fact that previous validation can be cited has now been spelled out more clearly (Antibody Validation section).

‘...information be mandatory when an antibody is used in a particular way for the first time’

The fact that if no previous validation has occurred then validation should be carried out and reported and/or submitted to a public database has been made clearer (Antibody Validation section). These are two key points and we appreciate the fact you raised them.

Things we have now included to respond to the numbered points raised.

  1. More discussion of the importance of knowing the antigen for an antibody has been added, in particular raising the point that for some antibodies the antigen is not known, for example when they are raised to a complex cell or tissue lysate (key details for reporting antibody experiments section).  

  1. This comment is relevant to the experimental validation of antibodies, we have increased the amount we cover on this topic but not attempted a full review as we feel such a complex topic is beyond the scope of this comment article. We have added some addition citations for readers who require more information (Antibody Validation section).

  1. We have made it clearer when validation should be carried out and how it should be reported if no previous validation has taken place (Antibody Validation section).

  1. We have now repeated the importance of including catalogue numbers in the antibody validation section.  

  1. This is now made clear (Antibody Validation section).  

  1. We think giving examples of available antibody databases will be useful to readers and were careful to mention more than one database, we have now added a link to a more extensive list. We have also removed the second reference to CiteAb which was in the final section.

Articles from F1000Research are provided here courtesy of F1000 Research Ltd

RESOURCES