ChIP-seq can only really work if you have a good antibody for the protein of interest. Any non-specific binding is going to add noise to your data making motif finding, differential binding analysis difficult or even impossible. Back in 2008 Mathias Uhlén's group at the Royal Institute of Technology, Sweden, published a study that showed how less than 50% of antibodies tested were specific to the target.
In February this year Andrew Bradbury and Andreas Plückthun wrote a comment in Nature on the problems this causes. Bad antibodies were blamed for the very high failure rate in replication of published studies, and the authors estimated some US$350 million was being wasted in the United States each year. They called for international collaboration and funding to validate antibodies through sequencing, and the use of only recombinant antibodies. Whilst this is a great start the sharing of data on antibodies by users is likely to go a long way to highlighting those that work and those that do not.
AbCam to the rescue - sort of: AbCam are now introducing knockout validation for antibodies as a standard QC procedure. It is likely to take some time to get their whole back-catalogue validated this way but if people buy more antibodies from AbCam than their competitors market forces will push others into improving validation processes. Ultimately we spend a lot of cash on antibodies, certainly the number of ChIP-seq libraries run through my lab would suggest an awful lot is being gobbled up to understand genome dynamics.
AbCam are using the CRISPR/Cas9-based human knockout library to verify that their antibodies are specific. This method guarantees there is no expression of the target gene/protein and confirms antibody specificity pretty well compared to other methods. They have some very clear gel images showing what sort of results are obtainable from "specific" to "non-specific" (see image above), and also provide a table below help interpret the results, what we are doing about it and what it means to you.
- Specific: Antibody is specific to the target only. It produces a band at the correct molecular weight, which disappears in the KO cell line.
- Partially specific: Antibody is specific to the target but also to other related proteins. It produces a band at the correct molecular weight, which disappears in KO cell line. However, extra bands are also present in both WT and KO samples. This may be due to cross-reactivity with other family members/isoforms or unrelated proteins.
- Non-specific: Antibody detects the target but shows a weak signal in the KO cell line. There is a band at the expected molecular weight in the KO cell line but at lower intensity compared to WT sample. There may also be some cross-reactivity with other family members/isoforms which are present at the same molecular weight as the target protein. Additional data and publication references will be provided to support this persistent band in the KO sample. If the non-specific band is identified as an unrelated protein or is without any supporting information, the antibody will be unpublished.
- Non-specific**: Antibody doesn’t detect the target. The expected molecular weight band is still present in KO at same intensity as WT. This antibody is not specific to the target. An alternate product which recognises the target would be recommended.
Abcam is aiming to validate 500 protein targets a year using this knockout approach as part of their antibody validation. Look for the logo below for knockout validated products.