Using  ideas and methods from linguistics to understand the language ("code") of proteins is something that I tried out 10 years ago ( M Gross, FEBS Lett., 1996, 390, 249), but somehow I never got round to investigate this more deeply or turn it into something really useful.

Researchers at MIT have now used peptide linguistics to do something extremely useful, namely to design novel antimicrobial peptides, making sure that they are highly active (because they obey the grammatical rules of other AMPs) but at the same time sufficiently different from natural AMPs to avoid the risk of blunting our natural defence.

Read my story here:

http://www.rsc.org/chemistryworld/News/2006/October/18100605.asp