Literature

"The genetic code is the mapping of 64 three-letter codons to 20 amino-acids and a stop signal." It stands out among possible competing codes in several ways. "First, the assignment of amino acids to codons appears to be optimal for minimizing the effect of translational misread errors." Errors in misreading a codon tend to have minimal effects on the translated protein. "Second, amino acids with simple chemical structure tend to have more codons assigned to them", as "they are required more often in protein assembly". But the researchers main interest in the paper below was the ability of the genetic code to carry parallel messages. The list is already impressive (and there is no reason why it should not be extended with research) - binding sequences of regulatory proteins that bind within coding regions, splicing signals that include specific 6-8 bp sequences within coding regions and mRNA secondary structure signals - all higher-order codes that ride over the protein forming code. "They found that the real genetic code could accommodate more arbitrary motifs in coding sequence than almost any of the other possibilities - it has a higher information content. One reason for the real genetic code's superiority is the fact that its stop codons, when frame-shifted, tend to form common codons, whereas in other codes frame-shifted stop codons form rarer codons or even other stop codons."
The authors appeal to selection to explain why the genetic code is optimal. The implication of this approach is that the selection had to take place before the Last Common Ancestor emerged on Earth. All this complexity had to be fine tuned in a single celled organism that predated all subsequent diversity. An information-based approach linked to Intelligent Agency deserves a fair hearing when seeking an explanation for optimal design.

The genetic code is nearly optimal for allowing additional information within protein-coding sequences
Shalev Itzkovitz and Uri Alon
Genome Research, 2007 17: 405-412. doi 10.1101/gr.5987307(Open Access)

DNA sequences that code for proteins need to convey, in addition to the protein-coding information, several different signals at the same time. These "parallel codes" include binding sequences for regulatory and structural proteins, signals for splicing, and RNA secondary structure. Here, we show that the universal genetic code can efficiently carry arbitrary parallel codes much better than the vast majority of other possible genetic codes. This property is related to the identity of the stop codons. We find that the ability to support parallel codes is strongly tied to another useful property of the genetic code—minimization of the effects of frame-shift translation errors. Whereas many of the known regulatory codes reside in nontranslated regions of the genome, the present findings suggest that protein-coding regions can readily carry abundant additional information.

See also:
Goymer, P., Evolution: The genetic code sees off rivals, Nature Reviews Genetics 8, 168-169 (March 2007) | doi:10.1038/nrg2076

There are many possible three-letter genetic codes that could adequately encode protein sequences, but what about the need to encode higher-order information on binding and splicing sites? New research shows that the actual genetic code is better than potential alternatives at encoding such information at the same time as encoding protein. [snip]

Evolution and multilevel optimization of the genetic code
Tobias Bollenbach, Kalin Vetsigian, and Roy Kishony
Genome Research, 2007 17: 401-404. doi 10.1101/gr.6144007

Abstract: The discovery of the genetic code was one of the most important advances of modern biology. But there is more to a DNA code than protein sequence; DNA carries signals for splicing, localization, folding, and regulation that are often embedded within the protein-coding sequence. In this issue, Itzkovitz and Alon show that the specific 64-to-20 mapping found in the genetic code may have been optimized for permitting protein-coding regions to carry this extra information and suggest that this property may have evolved as a side benefit of selection to minimize the negative effects of frameshift errors.

Last paragraph: "As we learn more about the functions of the genetic code, it becomes ever clearer that the degeneracy in the genetic code is not exploited in such a way as to optimize one function, but rather to optimize a combination of several different functions simultaneously. Looking deeper into the structure of the code, we wonder what other remarkable properties it may bear. While our understanding of the genetic code has increased substantially over the last decades, it seems that exciting discoveries are waiting to be made."