Teams now dominate in the production of papers and particularly of high-impact papers, according to a study of nearly 20 million papers by Wuchty et al. (Science 316, 1036–1039; 2007). This fact will be immediately obvious when you look through the table of contents of this issue. Success in the field of genome-wide association studies for common and complex diseases and traits needs large study populations and multiple replications. The author groups of these studies comprise formal consortia, consortia of consortia and project-specific collaborations among groups that have found their research moving along parallel lines.

When work is submitted for publication, it is journal policy for editors to communicate with the designated corresponding authors and not to discuss the manuscript (or even its existence) with individuals other than the authors, the team of editors at the journal and designated referees.

It is therefore the duty of corresponding authors to ensure that all authors on the paper have read and approved the submitted work. Citation or use of unpublished work from individuals other than authors requires a letter of consent from the data producers unless the data are already available in a public repository. It is therefore very important that the roles of authors, data producers and contributors be decided and detailed at the initial submission stage. Editors are trained to keep track of changes to the lists of authors and contributors during the review process, as often referees may demand more data. We may contact all listed authors and contributors if necessary. Referees increasingly want to see not only all contributors, but the data itself, so access to data sources and appropriate accession codes should be included before submission.

As collaboration speeds the pace of discovery, so simultaneous discovery becomes more frequent. In 2008, the journal has repeatedly handled the review and publication of related papers in concert. Where the simultaneous discovery is to be published in a competing journal, we respond by expediting publication of the paper accepted at this journal. We are a journal of record, publishing not the accepted author version but the fully edited, complete article with metadata, correct nomenclature and accession codes to our highest standards. Consequently, the production process is the work of teams of tens of individuals. You can work with us to expedite your accepted paper in both normal and fast-track publication by understanding how we handle your work and what can throw us off track.

The revised manuscript that we formally accept should be considered definitive. If time permits, the journal's copy editor will then often send the corresponding author an edited pre-proof incorporating the scientific clarifications of the manuscript editor and the English and technical edits of the copy editor. Revision—after formal acceptance—of display items, references, funding acknowledgments and, above all, author names, affiliations and contributions all pose recurrent problems that can delay publication. Page proofs should be used only to correct errors and proof turnaround times nowadays range from a couple of days to a couple of hours.

There are several places in the life of a manuscript where changes can and should be made. The best of these is the point at which your paper is accepted in principle. At this point it is imperative that all authors and contributors to the paper be contacted by the corresponding author to ensure that affiliations, contributions and funding sources are correct and up to date. All contributors should at this point update their affiliations in the journal's online database (where they have the opportunity to submit their own papers and to be considered as peer referees).

Scientific productivity is rapidly increasing thanks to collaboration, and this has created a greater need for communication and coordination. In response, publishers have now begun to offer researchers customized unique contributor identification services such as ResearcherID (http://www.researcherid.com/). However, we recognize that it would be unrealistic to expect something as central as individual identity and reputation to be definitively provided or controlled by any organization. What is really needed is a database or convention of online contributor identity, controlled by knowledge producers themselves, a service that records consortium membership with dates of joining and leaving, roles within consortia and authoring groups, and funding sources.