The search for a grammatical theory of gene regulation is formally justified by showing the inadequacy of context-free grammars.

No one questions the important practical contributions of computer sciences to molecular biology. It may well be that one day theoretical contributions also will become useful. One example of this type of interdisciplinary research is the attempt to construct a grammatical theory of the regulation of gene expression. In this paper, I demonstrate that context-free grammars are inadequate for the description of regulatory properties coded in the DNA. This result is supported by data available in the literature that show changes in the specificity of the recognition between regulatory proteins and their DNA targets. This result is an important limitation for the use of statistical approaches such as information theory as a source of inspiration for a theory of gene regulation. Additionally, such a demonstration gives formal justification to the search for more elaborate grammatical models in the study of gene regulation. Some basic proposals for such grammatical approach have been presented previously.