Theoretical Considerations
Gene sequences are commonly compared as two-dimensional alignments. It is useful to remember that significant homology between two sequences (DNA or deduced amino acid) implies general homology between their three-dimensional structures. Regions of homology within genes typically represent conserved structural features with similar relative orientations in three-dimensional space. In cases where structural information is available, the common way of displaying such information is to compare the fold, or Cα-carbon chains, from different proteins superimposed in such a way as to maximize superposition. There are thought to be ~1000 protein folds, at least an order of magnitude fewer folds than the number of enzymes (Zhang and Delisi, 1998). Typically, when the derived amino acid sequence homology is 25% or greater, the protein folds of two enzymes are likely to be very similar (Hobohm and Sander, 1995). However, there are cases in which the amino acid homology is too low to be detected by computer algorithms but the fold is highly conserved.