Ok, so I know the main conversation petered out a couple of days ago, but I've been keeping up on the thread since it started and doing some research since then. I also hold a BS in molecular biology with a minor in bioinformatics, and CRISPR/CAS9 has been a special area of focus for me over the time I spent in university, which means I should probably pitch in my two cents here, especially since this is such a hot topic.
I guess I'll start off by saying that I view this system as the equivalent of the development of high-grade steel at the beginning of the industrial revolution. We now have access to the most versatile 'material' in biology, but barely have any idea what to do with it. This is where bioinformatics comes in, which I view to be the equivalent of standard metrics, also historically developed around the same time. This is to say that bioinformatics provides us with an information system that is easy to understand and easy to manipulate, and above all, is standardized between all researchers. Ultimately I think this means that its almost literally impossible to predict what will happen in the future because of these developments. No one could have predicted the world of 2016 from what happened in the late 1700s, and the change in what my professors liked to call the 'genetic revolution' is likely to be orders of magnitude faster than its industrial counterpart. That's not to say we shouldn't speculate on the future, but we should do so with a
grain bucket of salt. There have been thousands if not millions of doomsday predictions based on new technology since the beginning of the industrial revolution, but not a single one of them has come to pass. There will likely be many more doomsday predictions based on genetics(and there already are), but like all the others, I think not one of them is actually likely to happen.
The other implication that this analogy has is that we really have very little idea how to properly utilize this system other than to collect more information about how our tools and the systems they operate on work. Scientists know how to use CRISPR/CAS9, but usually in very simple respects, such as engineering
E. coli that fluoresce using a green fluorescence protein(GFP), a common undergraduate experiment, or to delete single genes to create mutant models of organisms in order to understand various genetic interactions. The knowledge to do more technically does exist, but it is not fully understood, and the sad reality is that cultural stigma against science of this kind means that research will be hamstrung and limping pitifully forward until we can get over these mental blocks. Human embryo editing is universally illegal except in the very earliest stages of development, and those embryos cannot be carried to term or be allowed to develop. Even
en vivo (gene editing in a living organism) modification in humans is almost entirely illegal, with what few gene therapies(currently a very imperfect form of gene editing) that do exist being tightly controlled. Even non-human gene editing is tightly controlled and restricted. And everyone knows about the anti-GMO movement. Im not saying that we should go full mad scientist, but currently its so tightly controlled as to make it nearly impossible to progress. There are some notable workarounds, including the leukemia treatment mentioned above, but laws still restrict research so that this knowledge can't be developed to benefit humanity as a whole. So none of us can expect to see the benefits of this research for a very long time, or at least until people are actually educated on the issue and can improve their understanding. I actually have been called a servant of the devil before for telling someone I was studying genetics, which speaks volumes as to how the public looks at the issue, at least where I live.
That XKCD comic hit the nail on the head pretty well IMO, but it honestly only scratches the surface. That's kinda why we need so much computational power to solve biological problems, its simply too much information and pattern recognition for humans to do, no matter what. And I believe that many people misunderstand how much we know about cellular processes, without even involving multicellular organisms, cell to cell communication, tissue organization, intercellular protein scaffoldings, etc. Major new discoveries in this field continue to be made on a yearly basis. For example, in 2013, something called a nuclear pore complex(NPC) was studied in detail for the first time, revealing that its actually the tip of the iceberg in a system that recognizes and regulates the moment of DNA and RNA to and from the nucleus, based on extremely specific sequences that are continually reprogrammed by the cell. To put it bluntly it turned the entire fields of gene expression and viral defense on their heads, just three years ago. People say that we almost fully understand cellular processes, but in reality we only fully understand about half of them, and that's ignoring the processes involved when cells have to cooperate, which we have comparatively very little understanding of.
I'm currently watching the Luca Cardelli lecture, about halfway through. It seems to me like he's explaining how proteins already work to non-biologists. He does put forward some edited and formalized structures that force proteins to behave "exclusively" like logic gates. I put exclusively in quotes because protein interactions, especially the "switch" interactions that Cardelli talks about, work on a mechanism of protein concentration gradients. What this means is essentially that its technically impossible to have a protein give a binary output or receive a binary input. The behavior of proteins is usually read as what is called a sigma curve(the line graph looks like the greek letter sigma), which by definition has a slope and cannot be binary. I do grant that the chemical energy of protein interactions can be modified to make the slope extremely steep and mimic a binary input/output, but they can't be binary, very strictly speaking. The same goes for gene expression, since its controlled by protein interactions. As far as creating an RNA computer, I would compare the level of technology we have now to the first transistor. There was a bit of time, and quite a few scientific and engineering breakthroughs before transistors and all the other parts of a computer were able to be put together as a single working unit. Much more time and exponentially more breakthroughs were required to get us to the relatively 'smart' computers of today. While the idea of an RNA computer is fascinating, I doubt that we have the knowledge to make it feasible in the near future. Not to mention you would have to code CRISPR/CAS9 to be able to simulate electrochemical tendencies and organic reactions in order to custom-write RNA sequences to achieve a specific goal, which is something that traditional computers and models still struggle with.
As for RNA acting as proteins, RNA served as the main vehicle for chemical interactions before ribosomes and the genetic code of DNA were fully evolved to support a protein based chemistry, and continue to fulfill those functions on a limited basis in eukaryotes, prokaryotes, and archaea. tRNA(transfer), which is responsible for chaperoning the correct amino acid to the peptide chain based on the mRNA(messenger) codon presented. This goes without mentioning the ribosome itself, which is mostly comprised of rRNA(ribosome) machinery in addition to a few peripheral proteins. There are hundred if not thousands of more examples, but I'm sure you're all smart enough get the picture. There's also about another dozen classifications for RNA types that are essential in genetic regulation and other functions as well, including miRNA and wiRNA in case anyone wishes to look it up.
I think what I'm trying to say here is that the genome and its peripheries already do serve as a computer, even if it is one that was naturally created vs artificially. However, it is one that can change its structure and programming at will, unlike circuit based computers. What Cardelli proposes is simply cleaning up the "code" to make the biological computer more efficient. Simplifying code, removing unnecessary redundancies, making sure all the "hooks"(transcriptional modification proteins) match up to the proper "classes"(gene operons), etc. I'd be very surprised if the creation of an RNA computer were achieved by creating new systems rather than tweaking the ones that already exist. Ultimately I think this will be done based on a DNA code rather than an RNA code, since DNA is much more stable and immutable than RNA, and already has many mechanisms for modification by both RNA and proteins. RNA would most likely degrade too quickly on a chemical level to be useful for storing information long term, such as you might want on a functional computer.
LupusExMachina wrote:snip
Unfortunately the articles you posted were inaccessible without an account, but I can say that specificity(how identical a protein or sequence needs to be to interact with its intended target) is a major issue in genetic modification and even just in studying previously existing gene interactions. Unfortunately there's not really a way around it as far as I know, professionals simply have to do the tedious math in studying the chemistry of the molecules in order to increase the specificity towards the target sequence. I was able to find another article on the variant you mentioned, and it seems like that's how it achieves its goal, although it does it on a semi-systematic level with this variant, which is definitely convenient. You're right that it's nowhere near accurate enough for daily/industrial levels of use, and any variant would likely need to be orders of magnitude more accurate before that could happen.
And you're right about danger in the future too, although our universal policy of being overcautious in these fields is hampering progress. I think that a more suitable pursuit in this day and age would be to educate the public on these issues. After all, one cannot achieve the wisdom of when to use knowledge without first being aware of the knowledge itself and the application of it. Caution can be a good thing, but it shouldn't preclude the ability to progress, as it is doing now to a large extent. I don't want to extend this into politics, I just want to make the general statement that an educated opinion is better than an uneducated one, and is important to have in this day and age.
EDIT: grammar