Linear and nonlinear constructions of DNA codes with Hamming distance d, constant GC-content and a reverse-complement constraint

In a previous paper, the authors used cyclic and extended cyclic constructions to obtain codes over an alphabet {A,C,G,T} satisfying a Hamming distance constraint and a GC content constraint. These codes are applicable to the design of synthetic DNA strands used in DNA microarrays, as DNA tags in chemical libraries and in DNA computing. The GC-content constraint specifies that a fixed number of positions are G or C in each codeword, which ensures uniform melting temperatures. The Hamming distance constraint is a step towards avoiding unwanted hybridizations. This approach extended the pioneering work of Gaborit and King. In the current paper, another constraint known as a reverse-complement constraint is added to further prevent unwanted hybridizations. Many new best codes are obtained, and are reproducible from the information presented here. The reverse-complement constraint is handled by searching for an involution with 0 or 1 fixed points, as first done by Gaborit and King. Linear codes and additive codes over GF(4) and their cosets are considered, as well as shortenings of these codes. In the additive case, codes obtained from two different mappings from GF(4) to {A,C,G,T} are considered.