The classification of transcription factors

The basic idea of this classification was to come up with a systematic of DNA-binding domains, in analogy with the well-known enzyme classification, and assign each TF a four-digit classification number, plus two more digits for its instance (gene & gene product). One of the aims has been to provide a clue to the protein-DNA recognition code, so that for a newly discovered TF, a prediction can be made with regard to their DNA-binding specificity just from the properties of its DNA-binding domain.

A comprehensive classification scheme for eukaryotic transcription factors (TFs) that is based on the properties of their DNA-binding domains has been published first in Wingender, Classification scheme of eukaryotic transcription factors, Mol. Biol. Engl. Tr. 31, 483-497 (1997) (full text here), and was later refined (Heinemeyer et al., Nucleic Acids Res. 27, 318-322, 1999; PubMed 9847216). It was closely linked to the TRANSFAC® database. For details of this previous classification, please see links given at the end of this page.

A more formalized approach to revise the classification and to assign newly discovered transcription factors to the established classes was published in 2004 (Stegmaier et al., Genome Inf. Ser. 15, 276-286, 2004; PubMed 15706513).

In the last years, however, a huge number of new types of DNA-binding domains have been found and have to be classified as well. It is therefore the aim of this project to enhance the previous classification scheme by introducing new classes, summarizing them into new superclasses whenever necessary, and to come up with an automatized classification tool. This new classification has now brought to a certain level of maturity for human TFs, so that the following resources have been made publicly available:


The following files document some features of the previous (and outdated!) TF classification, published first in 1997 (PubMed; Full text (pdf, 15MB)):