Is there any library or application I can use which would allow me to write DNA and test how it runs on a virtual single celled organism? If not, what can I read that would allow me to learn how to create my own DNA interpreter? I know they have a programming language that can be used to compile DNA sequendes for bacteria, but I want to learn the language of raw DNA.
DNA simulator?
Noah Garcia
Robert Wright
>sequences
fixed
Ayden Carter
I'm pretty sure I remember seeing something along those lines in the Ubuntu repos.
Adrian Gonzalez
Can I compile a waifu?!
Wyatt Clark
I would first try to graduate high school. This should give you enough understanding of biology to realise why your question is complete bollocks. Cheers!
Brandon Baker
You can do something like that but it will absolutely not be biological.
If you wanted to simulate something like that, and I don't know why you would, you could I guess generate a random string of A T C and G, then use some bioinformatics pipeline to guess at what it might do in an organism.
That would look something like using BLAST to search for homology with sequences on NCBI and then breaking it up into different sized units to figure out all the potential coding regions.
Again, don't do any of this, because it won't tell you anything about what it would do in an organism.
Juan Johnson
So do you completely deny the fact that people are currently able to use a programming language called Asimov to create DNA sequences that can be run on e-coli?
Jason Roberts
Yes, that's a hoax made to generate money
Jaxson Nelson
1) genetics are a hell of a lot more complex than "m-muh DNA sequence"
2) computer simulations of protein folding never gives correct results
3) molecular docking simulations for a single protein and a single small organic molecule literally requires a supercomputer, you can forget any kind of simulation involving hundreds of millions of proteins and compounds all interacting simultaneously
4) about 10 billion other reasons why you're a retard that I'm not going to get into
Ayden Evans
Proof?
Alexander Anderson
>molecular docking simulations for a single protein and a single small organic molecule literally requires a supercomputer
I have about 50 reasonably powerful GPUs at my disposal (everything is either a 1060, 1070, or 1080ti). Think that would be enough?
Jace Nelson
...
Jayden Reed
>programming isn't technology
retard detected
Charles Morales
That's an ill-posed question. You cannot ``run [a DNA sequence] on e-coli'', that's simply not how it works. And the guys from Asimov have craftily not shown any proof of concept, just some science fiction wish-wash. Not to mention that the idea seems pretty ridiculous, as they are not accounting for any of the cell's epigenetic dynamics. It's piss easy to create a simple gene regulatory network, but that's only a miniscule part of an extremely complex system, even for a bacteria like e.coli.
Wyatt Sullivan
No.
Spend your electricity bucks on something worthwhile instead.
Like a good sex robot.
Ryan Allen
free electricity (although I do pay for rent)
Oliver Young
That's about 500 Teraflops if we're being generous.
book.bionumbers.org
This site gives us an estimation of about 2-4 million proteins for an e. coli bacteria, lets go with 2 million, at an average of 300 amino acids per protein. The mean number of molecules per amino acid is about 19, so that would make for at most 2,000,000 * 300 * 19 atom interactions for each timestep. Obviously it would be lower than that, as they aren't interacting with each other all the time. Lets again be generous and say 1/10th of the molecules are interacting at any given time, that's still more than 1 billion molecules that need to be simulated.
en.wikipedia.org
Lets pretend there is software that is smart about what needs to be simulated and what doesn't at any point in time, I don't know if such software exists. At best, algorithms doing that have a run time of O(n * log(n)) for n molecules, and lets finally pretend the software needs 1k FLOPs for a single interaction. Usual time steps for such systems are one femtosecond.
If you'd want to simulate the bacteria for 1 time step with 500 Tflops, given all these constraints, it would take you about
0.018 seconds. To simulate it for 1 real time second, it would take you about 1.8 * 10 ^ 13 seconds, or 18 trillion seconds, or about 570,000 years. Good luck m8
Wyatt Watson
Simulations don't always need to be perfectly accurate to the lowest level possible. Surely, once something is understood mathematically, it becomes possible to take shortcuts, as has been seen in neural network simulations which do not even come close to actually simulating the low level functionality of a neuron.
Adrian Walker
This already takes a shortcut, using an O(n * log(n) algorithm, instead of the usual O(n^2).
Your analogy is flawed, unlike with neurological systems, where your desired inputs and outputs are simple electrical signals, where you don't care what happens in between, as long as it behaves correctly, this problem requires actual physical simulation of everything to give a desired output, which is the emergent physical properties of simple molecules forming a biological system. You can't compare these two, and there aren't a lot of shortcuts to take until quantum physics are solved.
Asher Lewis
No. But you can simulate miR binding on DNA targets with mirbooking nowadays.
Most gene functions aren't known yet (see every gene that starts with Gm, or in general check out relevant GO terms for genes) so you wouldn't be able to write a DNA interpreter anyway.
Yes, that doesn't exist.
Daniel Scott
Juat cum inside of your usb port.
Aiden Moore
HAHAHAHAHAHAHAHAHAHAHA, oh god no.
>> virtual single celled organism
we are so fucking far from that it's not funny. Right now it takes incredibly HUGE amounts of computational power to simulate an entire virus for a couple of microseconds. Even then that's at a lower resolution than all atoms, and recent results suggest that we actually do need to simulate all atoms. All this just to model the virus flopping around.
that's cute. You need processors more than you need GPUs for molecular dynamics. We're talking months of sim time on at the very least on of the top 5 of the top500 supercomputers for the virus simulation.
Simulating entire cells is currently beyond our computing technology. It's basically impossible to do much more than watching stuff flop around passively, because of the huge timescales of phenomena we are interested in relative to the timesteps we must use. To model a bacteria we need to be able to simulate at the seconds timescale, but we have to use nanosecond timesteps for MD to work, femptoseconds if we want to simulate all atoms. If you do the math, the amount of timesteps you have to simulate to reach the seconds timescale is mindbogglingly TITANIC. Like if you started the simulation today, it wouldn't finish in a millenia. We'd need vastly more powerful computers to simulate stuff like this.
Jaxson Morris
>recent results suggest that we actually do need to simulate all atoms
Can you please elaborate on this or provide me with some reading material? That sounds fascinating. Also, what kind of degree to I need to get a job where I can play with this kind of tech?
Gabriel Bennett
this was not the whole virus, just the virus capsid:
nature.com
It's an ABSOLUTE UNIT of a simulation though. They simulated for 1.2 microseconds with 2 fs timesteps. They used 3880 GPU accelerated Cray-XK nodes on the Oak Ridge Titan to do this.
>>Also, what kind of degree
a PHd chemistry, physics, biology, or materials science focused on doing really big molecular dynamics simulations. Find a group that does this work and tailor your degree to get into that group and do a PHd with them.