When Nature or its sister journals receive serious allegations about data or author conduct, they follow a clear procedure to work out whether the published record needs to be revised