Health-care systems spend billions of dollars annually on biomarker research for personalized medicine. Success hinges on the quality of the biobank specimens and the data used to derive them, but a lack of quality control is polluting the scientific literature with flawed information that will take a long time to sort out (see http://go.nature.com/tfm8sn).

We analysed 125 papers (see http://go.nature.com/yagpxx) retrieved in a PubMed search of open-access articles using the key words 'biomarker discovery' for the years 2004 and 2009. More than half of these contained no information about the biospecimens used. Four papers on biomarker discoveries published in Nature in 2009 contained insufficient specimen data.

Leading journals are trendsetters when it comes to defining publication criteria. For example, for some 15 years they have required statements on ethical review boards and informed consent; a biomedical paper without this information would be unthinkable today. To uphold standards, all journals should insist on full details of biobanked specimens (including medication, transportation, processing and storage).

Thousands of potential biomarkers are reported every year. Biobankers should collect complete information on specimens and pass it on to researchers to include in their publications.