Neural Network Assisted Evolution

Expanded DNA alphabet provides more options for nanotechnology:

Long-time readers of Nanodot may remember the section of Chapter 15 of Nanosystems in which Drexler explores options for producing easier to design proteins for the protein engineering path toward atomically precise manufacturing by incorporating specially chosen amino acids in addition to the 20 genetically encoded amino acids. Back in 1992 the only option for incorporating unnatural amino acids into proteins was Merrifield solid phase peptide synthesis, using the methods of organic chemistry rather than biological systems. However, this becomes problematic and expensive for longer chains. Consequently, finding ways to expand the repertoire of biologically encoded amino acids would be quite useful. One way to accomplish this goal would be to expand the DNA ‘alphabet’ from two to three base pairs (that is, from four to six ‘letters’). We noted progress in this direction back in February of 2008 when Floyd Romesberg, at the Scripps Research Institute, La Jolla, California created two artificial DNA letters that were accurately and efficiently replicated by a natural enzyme. In September of 2011 we noted a different approach taken by a team at the Salk Institute that keeps the current DNA alphabet but alters one three-letter word to mean an unnatural amino acid, increasing the amino acid repertoire by one. We noted in June of 2012 that continued work by Romesberg had revealed how the new base pair was efficiently replicated in the test tube by a natural enzyme. In a major advance, Romesberg and his collaborators have engineered a living organism to stably propagate the expanded genetic alphabet...