There are several challenges that must be overcome to validate the specificity and reproducibility of antibody reagents and this has been emphasized by many recent publications. One important issue is cross-reactivity due to off-target binding, which is defined as antibodies binding to proteins other than the intended target. Another issue is that samples are treated differently in different applications, which influence the epitopes exposed on the target protein. As a result, this might have profound consequences for the ability of a given antibody to bind specifically to its target. Thus, antibodies must be validated in an application-specific manner, as was recently pointed out by the International Working Group for Antibody Validation (IWGAV). The working group proposed five pillars for antibody validation, all allowing the antibody to be validated without the need for any prior knowledge of the target protein, except the gene and protein sequence. This is important since many of the proteins predicted from the genome sequence lack previous literature and they are thus important targets for antibody-based studies to allow a genome-wide analysis of the entire human proteome. Here, we have adapted these pillars for Western blot applications with a focus on providing scalable and streamlined methods. We present a strategy for enhanced validation suitable for commercial providers as well as small sized research groups with limited resources.
The Western blot application is the most frequently used antibody-based method, with approximately 1.5 million antibodies classified as supported for this application in the Antibodypedia portal. In a Western blot assay, the approximate size of the target protein is obtained as part of the analysis and off-target binding can be probed by the presence or absence of additional bands. However, many protein bands are shifted in size compared to the predicted molecular weight, e.g. due to proteolytic processing and various post-translational modifications, including glycosylation. Yang et al. cataloged such variation by carrying out proteome-wide quantitative mass spectrometry (MS) in a vast number of slices from polyacrylamide gels and reconstructing virtual Western blots that were compared to predicted molecular weights of unmodified proteins. Around 15% of the proteins had their most prominent band far from the predicted molecular weight and even more had multiple bands that all mapped to the same protein-coding gene. Consequently, there is an obvious need for validation principles that are independent and complementary to the theoretical size estimate used in the standard Western blot assay.
Here, we show that the methods described can be used for streamlined validation of antibodies for Western blot applications using convenient panels of cell lines for the analysis. More than 6,000 antibodies were validated in at least one of the five methods and all the primary data for the validation is presented as part of the Human Protein Atlas (www.proteinatlas.org) antibody info page for each respective antibody A special focus was to investigate the performance of the orthogonal and gel-migration capture MS validation strategies, since there are no prior examples in the literature of systematic validation of antibodies using these two rationales. In the orthogonal validation strategy, the protein abundance levels obtained using an antibody-dependent method are compared with the levels determined by an antibody-independent method across a set of samples. For the capture MS strategy, the apparent size obtained by an antibody is compared to the presence of MS-determined target peptides after cutting out gel slices. Altogether 1,630 antibodies were validated by at least two of the pillars and 267 were validated by three or more pillars. The results show a path forward for streamlined validation of antibodies, although it is important to point out that the enhanced validation is specific for a certain sample context and the validation is thus dependent on the sample preparation procedures used to evaluate the assay including the relative abundance of the target protein.