[Tyler]

Coronavirus Genome 2
27 April 2020, 5:00 am



Source: xkcd.com

[ivan.moony]

It's in attachment.

[edit]
Now, we just need to find an anti protein that brakes it.

[HS]

Yup, are you still going on project Rosetta? I've currently gotten to 85,907 with a core i5 cpu.  Any other methods out there?

[ivan.moony]

Yup, are you still going on project Rosetta? I've currently gotten to 85,907 with a core i5 cpu.  Any other methods out there?

Great score, you're the man over here. With me something went wrong, I got to about 12,000, then it stopped working, some error about connection, it says. I should try to fix it more carefully.

About other methods, there was some protein folding turned into collaborative online game where human players compete to manually construct the least energy setup protein they can. There was high scores, hall of fame, hang about forum, and everything else. The project name is FoldIt.

I also liked PV project, if you are after building your own protein viewer app. And here is some aminoacids reference (they compose proteins that compose organisms including viruses). There is surprisingly low number of possible aminoacids (about 20), yet, combined, they can compose all the variety of life on the Earth. Amazing stuff.

I also linked a Tyler's announcement MIT article on using AI on protein folding prediction. It is neural networks stuff with genetic algorithm, if I recollect correctly.

Some time ago, I was after polynomial solution to traveling salesman problem that would help in speeding up protein folding prediction, but no success here (I think I got only O(2ⁿ) solution, which is exponential and not good enough achievement). Success would mean O(n^k) solution, reducing time from computer decades to computer minutes, or so.

[HS]

About other methods, there was some protein folding turned into collaborative online game where human players compete to manually construct the least energy setup protein they can. There was high scores, hall of fame, hang about forum, and everything else. The project name is FoldIt.

Whoa! I didn’t know about any of this stuff. They are Ender’s Gaming it! If this could be put into a mobile game app it could reach a lot of young people with probably unbelievable potential to get good at this. 

There is surprisingly low number of possible aminoacids (about 20), yet, combined, they can compose all the variety of life on the Earth. Amazing stuff.

There's some type of thought process that humans are obviously missing to be able to intuit this type of thing.

Some time ago, I was after polynomial solution to traveling salesman problem that would help in speeding up protein folding prediction, but no success here (I think I got only O(2ⁿ) solution, which is exponential and not good enough achievement).

You might think it's not very good, but you always impress me with this stuff. I had to watch YouTube video explaining big O notation...

[ivan.moony]

About other methods, there was some protein folding turned into collaborative online game where human players compete to manually construct the least energy setup protein they can. There was high scores, hall of fame, hang about forum, and everything else. The project name is FoldIt.

Whoa! I didn’t know about any of this stuff. They are Ender’s Gaming it! If this could be put into a mobile game app it could reach a lot of young people with probably unbelievable potential to get good at this. 

It should be a great feeling, being a part of such a community gathered around something that much important like fighting diseases. Incredible things are possible when people are connected.

I finally reinstalled Boinc manager, and all the account data was restored. Interestingly, things broke right after reaching 13k, but all of it was regenerated successfully, so tasks continued where they stopped. I'm aware that this effort is not much contributing the entire medical status, but if they say it counts, what a heck, I like being useful.