Only Good Antibodies — characterisation

Highlights antibodies cited on a page and shows the knockout-controlled characterisation data for the application they were used in.

Most antibodies in the literature have never been tested against a knockout. Only Good Antibodies marks the ones that have. ALL RESULTS DESCRIBE PERFORMANCE IN STANDARDISED CONDITIONS. Antibody performance is assay and sample dependent, so what you see is independent characterisation data, not a verdict on the antibody. The results in the tested conditions are a useful guide to the quality of the reagents a paper relies on — most useful where the paper provides no context-specific validation of its own. If a paper has used proper controls (a knockout or knockdown in the sample of interest), it may show that an antibody whose data was not supportive under the conditions we tested was still suitable for the conditions used in that paper. Open a paper on a journal, preprint server or Europe PMC, and every antibody in the methods that YCharOS has independently characterised is underlined in place, in the paper's own text. Hover it and you get how that antibody performed in each application — western blot, immunoprecipitation, immunofluorescence, flow cytometry — with the wild-type-versus-knockout image the result came from, the clone, the RRID, and a link to the published report. The result is reported per application, not per antibody. One antibody can carry three different answers in a single card: Supportive — characterisation data supports this application in the conditions tested. Limited support — the antibody detects its target, but support for this application is limited. The same antibody can be limited for western blot and supportive for immunoprecipitation and immunofluorescence. Not supportive — tested with knockout controls, and the data did not support this application. That may still mean the antibody detected its target but was not selective; the gene page says which. Two antibodies in one sentence can differ, and often do. Not tested — in the dataset, but this application was never tested. An antibody nobody has tested in an application is not an antibody that failed. No product identified — the paper names a target but gives no catalogue number, or this exact product is not in the dataset. The extension says so plainly and points to the antibodies for that target that do have knockout-controlled data. A toolbar panel summarises the whole page at once: how many antibodies on it are supportive, of limited support, not supportive, split across applications or untested in the application used — plus characterised alternatives. Each card also offers to hand the paper to your own AI assistant, through the OGA connector, to work out which application the paper actually used and whether it ran controls of its own. The data comes from YCharOS, an open-science consortium that characterises antibodies against knockout cell lines to a community consensus protocol and publishes every result, including the negative ones, as a peer-reviewed report with a DOI. Nothing here is a supplier claim or a citation count. How it handles your reading is the short part: matching happens entirely inside your browser. The extension downloads a small index of antibody records from onlygoodantibodies.co.uk once a day and compares it against the text of the page you are on. Nothing about the page, the paper or you is transmitted anywhere — there is no account, no analytics, and no server that sees what you read. It works on around a hundred publisher, preprint and repository domains out of the box, and covers institutional proxies and supplier catalogue pages if you turn those on in settings. Free.