Most scientists would agree that graduate school packs in an incredible amount of education. But do PhD programmes teach all the right things? Most PhDs would probably say that they learned a lot about science — from research methodology to literature reviews and experimentation. But there are many other, equally important, skills that often fall through the cracks. As science and technology continue to move on, this 'soft' skill set remains more or less constant.

Soft skills include learning how to pick mentors, write grant proposals, assemble interdisciplinary research collaborations and manage labs. Without mastery of them, meeting scientific goals and climbing the ladder towards tenure can be more difficult. Most young scientists scramble to pick these skills up on their own, when they realize they need them.

Last year, I talked to hundreds of graduate students and postdocs, and I read scores of reports about the need to improve graduate education and to speed young scientists towards working independently. As a result, Naturejobs ran a series of articles listing soft skills and showing how to acquire them. This week, that entire series, together with additional relevant articles, has been posted in the Editor's Choice part of the Naturejobs website (http://www.naturejobs.com).

I hope that this will encourage administrators to incorporate some of these skills into graduate education. But if they don't, then at least young researchers should be able to pick up what they need to know faster, on their own.

In an ideal world, administrators would not only ensure that graduate students get this sort of training, but that undergraduates are aware of the skills. Because undergraduates who join graduate school knowing how to pick mentors and speed through the system will have a more productive time focusing on science — making the most of what graduate school already does well.