The BRD have recently been shown to be involved in chromosome organisation and possibly chromatin remodelling during spermatogenesis (Shang, E. et al, 2004), and their initially reported kinase activity (Denis, G.V., and M.R. Green, 1996) has not been reproduced thereafter by independent groups (Debra Wolgemuth, personal communication). Proteins of the ABC1 group have never been biochemically characterised as protein kinases, although a certain similarity to subdomains I, II, VIb, and VII of the ePK catalytic domain has been noted (Leonard, C.J. et al, 1998). The TIF1 group were initially thought of as protein kinases (Fraser, R.A. et al, 1998), but the kinase activity has not been recognised in subsequent work by the same group (Khetchoumian, K. et al, 2004). H11 is exceptional in that it displays protein kinase activity, although it is neither an ePK nor an aPK, but a chaperonin, Hsp22 (Kim, M.V. et al, 2004). FASTK, initially characterised as an aPK in a Fas-apoptotic response (Tian, et al 1995) has finally been shown not to be a kinase (Li, W., et al, 2004). G11 was initially shown to display kinase activity (Gomez-Escobar, N.C.F. et al, 1998), but this activity could have been due to the protein binding a real kinase tightly since G11 was made in COS-7 cells followed by immunoprecipitation. Its kinase activity has not been characterised since. Two other proteins, BCR and TAF1, display real protein kinase activity, but have not been included for the generation of atypical-kinase-based HMMs for two different reasons. BCR remains a protein kinase in its own right, but is a special case of gene fusion with the Abl tyrosine kinase in chronic myologenous leukaemia (Hehlmann, R., Berger, U., and A. Hochhaus, 2005). TAF1, a component of the basal transcription machinery, has been shown to have kinase activity (Maile, T., Kwoczynski, S., Katzenberger, R.J., Wassarman, D.A., and F. Sauer, 2004), but its C-terminal kinase domain bears no resemblance with the ePK or aPK domains. However, since it possesses a very distinctive signature (InterPro 011177) and only one copy of the gene appears to be present per genome, it can be regarded as a special case that did not fit into the framework of ePK and aPK domains. A6, a single, standalone, aPK has so far failed to be biochemically characterised as a protein kinase, although it has the ability to bind ATP (Rohwer, A.W. et al, 1999).