Help save the world with Rust! I'm the security lead and software architect for the SecureDNA project, and we're looking for a development manager to help build a team and implement the specification we have developed. (If you are a software developer and would like to be notified when we advertise specific positions in the near future, feel free to drop us a line; I will also reply to this post, or make a new one, with specific positions.) These are 100% virtual positions, with the flexibility to hire as full-time salaried or as a contractor at competitive rates. Read the complete description of our open positions at the project website, and see below to whet your appetite.
The SecureDNA project seeks to enable fast, reliable, cheap, automated screening of DNA synthesis orders, to prevent both careless and malicious users from synthesizing potentially-hazardous DNA sequences anywhere in the world. It is an international, academic/commercial partnership project, consisting of researchers from MIT, Tsinghua University, Aarhus University, Shanghai Jiao Tong University, and other world-leading academic institutions, in partnership with major international DNA synthesis providers.
Our goal is to develop an automatic and accurate screening system which can:
Screen DNA synthesis orders much more quickly and cheaply than the current human-reliant review system;
Have negligible impact on legitimate research by researchers authorized to work with particular organisms;
Effectively block hazardous DNA sequences from being produced without suitable oversight, whether via accident or malice;
Protect intellectual property by not disclosing the submitted DNA orders to third parties;
Protect worldwide security by not disclosing potential hazards which may not be public; and
Enable global commercial adoption, with community verification, via scientific diplomacy and open-source software.
Join us! Use Rust's robustness and security to help make the world safer place.
Preventing human extinction by preventing known accidental spillage sounds like a nice doable goal.
In long run, we should somehow have ability to customize some set of interfaces/processes inside all our cells which creates incompatibility between human biology with anything that's trying to get translated/transcripted.
I'm a assuming you're thinking mainly of viruses, which at face value does seem to make sense.
But given that our own DNA and various kinds of human-native RNA also do that, it sounds to me like this has no chance of working, at least not as described.
In addition, there is evidence that certain codons in the human DNA didn't originate in what we generally consider our ancestors. Instead they originated in viruses (of all things) that incorporated themselves into the the DNA of our ancestors. Making viruses incompatible with humans would prevent things like that from happening again.
I am not sure why I would want a virus to embed itself in my DNA. Yes, creating incompatibility would prevent things like that from happening again, if it happened it in past. But we don't need random chance to improve our biology, we now have tools to experiment and improve our biology. One area being explored is making human biology adaptable to space.
Some background for the uninitiated:
DNA goes through transcription to produce RNA, RNA gets translated into Proteins by the Ribosome.
To my knowledge, all biology uses same set of mapping between codons and amino acids.
A simple solution:
All biology is using DNA in particular configuration, mirroring DNA introduces incompatibility with existing biology.
A complex solution:
We use a different mapping(inside ribosomes) betweens codons and amino acids, this means tweaking our ribosomes in the right place. This mapping can be unique to every other human.
Benefits? All existing RNA(and DNA) will stop making sense, it will produce random proteins if a virus injects its DNA/RNA into the cell.
Drawback? This also applies to us, even our own DNA will stop making sense, epigenome will be messed up.
Overcoming drawback? At time of cell division, we modify the Helicase protein to do this necessary adjustment to our DNA, so new DNA can be correctly converted into right proteins.
Other drawbacks? Inhibitors may bind to incorrect site location on DNA so they will also need to be adjusted. I don't know enough to comment on this.
Disclaimer: I am no cellular biologist, so this is based on my limited but a high level understanding.
Security in computer systems is pitiful. Everyday we read in the news of a new major data breach. Or ransom attack. And so on. Even staying off the net does not help, remember Stuxnent? The situation shows no sign of getting any better. Even governments are waking up to what a disaster this all is : Senate Democrats call on FTC to fix data privacy ‘crisis’ - The Verge. Never mind the threats to national security.
I conclude the probability of keeping dangerous DNA sequences out of the hands of bad actors is vanishingly small. It's just data in files, right? If the mightiest software enterprises on the planet can't solve this problem how can we assume anyone else can?
Take for example the placenta. It exists because of a retrovirus that literally causes some cells of the fetus to merge with cells of the mother.
Viruses mutate pretty quickly. I am sure that we will get a virus capable of infecting inverse dna humans in maybe a couple thousand years, if not much quicker. Also note that there is a many-to-one relationship between dna triplets and aminoacids, so it may be possible that a single virus gets a dna sequence that translates to similar enough proteins for normal and inverse dna organisms such that this won't be a barrier for viruses for all too long. It will however result in a significantly weaker immune system, so the moment a virus can jump over to inverse dna humans, it will likely be very deadly and wipe out a significant part of inverse dna humanity. Or maybe I am misunderstanding your proposal?
You are right. It should take nature quite some time to overcome this but a bad actor could enginner a virus in a lab by inversing RNA of a virus and viola we have a virus capabale of infecting humans with inverse dna sequence.
Mirroring DNA has some other challanges as well, we will also have to mirror DNA sequences of foods we eat, our bodies wont recognize existing ones because they will expect all proteins they interact with mirrored as well.
They like to tell you the genetic code is universal in high school biology, but it's been known for a while now that this isn't the whole story. There are even differences within one organism, including humans - mitochondria and nuclear mRNA is translated slightly differently, because mitochondria have retained the mapping from its prokaryotic past.
Ribosomes are cooler than that! They can handle arbitrary codons (mostly, there are complexities) - what determines the mapping between codons and amino acids are tRNAs, and in turn, a group of aminoacyltransferases called the aminoacyl tRNA synthetases. By expressing a different set of these enzymes, you could (and organisms have) changed the mapping of the genetic code.
However, you can't really arbitrarily mix up the mapping, as life has optimized it for millions of years to find a solution that works really really well. There are biochemical reasons why particular codons match to given amino acids.