The overall conclusion from the survey described here is that the majority of ARID subfamily domains bind DNA without regard to sequence specificity. Thus, the acronym is somewhat of a misnomer, although it is a well established and useful descriptor for a domain whose parameters are well-defined. This survey did not probe the behavior of every single member of the human ARID family. The proteins that have not been tested directly, here or elsewhere, among the subfamilies now designated as sequence non-specific are RBP1L1 (ARID4B), SMCX and SMCY (JARID1C and JARID1D). Each of these shows at least 75% identity and even greater similarity to the tested members of its subfamily. In addition, a mention of data not shown in a report on RBP1L1 (syn:SAP180) notes that a high-affinity consensus binding site could not be found in DNA-binding site selection experiments (25). Among the subfamilies now designated as AT-specific, only DRIL2 (ARID3B) and ARID3C have not been tested empirically, but again, there is at least 75% identity and more than 90% similarity between these ARID sequences and the AT-specific prototypes Bright
