GENE EDITING

latest link

My interest goes back to 1976 when I studied and liked L-systems (Lindenmeyer). Botanists were trying out L-systems to understand the construction of tangle of flowers. This requires propagation rules in context sensitive grammars. Context-free were the rage then, but insufficient to control, specific changes in several sites. This time dependent behavior required rewrite rules with context! Specific required context could be propagated to desired places only.

I became a compiler writer with expertise in yacc which dealt only with LR(1) subsystems of context-free languages only. I developed great attachment to attributes, rewrote yacc to allow attributes and come up with two level context-free (LR-1 really) and two-level LR-1 really is most powerful possible(two-level CFG = type 0)! Attributes are really for context-sensitive ness! When done, attributed LL-1 grammars will rule the world, even better than Go. Go is great for it encapsulates, Ken Thompson ideas on parallelism and has parallel garbage collection!  My language, improving Go is eventual, though more like Haskell.

But here we only apply to crispr for genetic editing. One can capture scientist work to create an appropraite model of production of an organism from the root cell. Despite ethical storm from germline edits, I limit to crops! There is a worry that in using CRISPR-Cas9 to repair one disease-causing mutation in human embryos, other potentially harmful mutations may be unintentionally introduced. To understand this consider context sensitive rewrite.

A  T  G C

G T C A existing DNA

T C A G

In this suppose out goal is replace middle sequence with AGTC  TAGC. OUR conext-sensitive rewrite is

A  T  G C         A  T  G C

G T C A  =>     A G T C

T C A G         T A G C

                    T C A G

Clever bio-chemistry later (context and replacement becomes an RNA that mixes with DNA, mix replaces this context sequence atgc tcag in all places. ATGC TCAG are matched and cut happens in middle there.

DNA forms a double helix. Ca9 may cut both, different RNA only 1, then two different RNA are needed. Both can have different contexts. One can improve cutting so. The chance of wrong cut is smaller! Together with more context, accuracy may be increased.

Crispr-cas9 is best now, but competes with transcription activator-like effector nucleases (TALENs) and (ZFNs).

Recently a Chinese scientist He applied to human germ-line, produced babies, and was roundly criticized! Easy bad science giving bad-name to non-germline research! Why? Only way to trace errors is repeated genetic testing of humans so produced! What if error found, what do you do! What if error in germline DNA? Horrible questions, no answers. Non-germline errors in one in individual only!