Skip to Main Content
The paper presents an exemplary computer representation of protein structure. The paper described how the configurations of protein backbone turns can be recovered from shape-string (sequence of symbols) descriptions, which are compressed and approximate, by looking them up in a table. Although they are used internally in many algorithms, shape strings are not generally regarded as an annotation that humans should read. Protein structure is much more conserved by evolution than sequence is. Shape strings can be very revealing to the human eye and should identify distant homologues whose sequences have diverged in the fog of time.