Well, will try to share my vision about this
First of all, it's obvious that fold proteins as PoW will be great, but it's obvious too that if it isn't done yet is because is very difficult. Some problems which come to my mind without thinking much about and some with possible solutions are:
- Time betwen blocks if we use a Bitcoin-like model (I think this could be solved if something symilar to eXocoin algorithm is used, which uses a timer for subbmit new blocks and once the block is solved can't be submited and miners becomes idle till new block need to be found) or at least this is which I understood about it without any technical knowledge
-Folding the protein only from their aa sequence will take toooooooo long, each aa have 10 liberty degrees in the space, so if a small protein will have 300aa, 10^300 are many possible combinations, this is solved using data about secondary structure and 3 and 9 nucleotida fragments structures for guide the tridimensional protein structure formation (more and more problems to add into the coin code which are solved nowdays using different applications which even uses databases to find this kind of information)
-The model always can become more accurate and will take many time, you have to say when you consider it valid, and this minimun energy state woudn't be universal for all proteins, and without knowing much about, I think this will be hard to explain to a computer which will use always the same code for fold all the proteins and decide when is really folded.
-The utility of ab-initio models isn't very huge in my opinion, they give you an aproximate model, but I think it will be more trustoworthy in many cases obtain it thrw x-ray cristalography and refine it . (This models could become more usefull for theorical proteins which don't even exist, but in many cases, this proteins, I think, derive from real proteins with some modifications)
Sure there are many more problems, but it's late here, I didn't slept enought yesterday and today I get very bored in class, so my brain is diying and don't want to work with me
So I think that one better alternative to have a coind distributed throw this method, is to use a POS system as base, for example, I said in other threads, use the nodecoin in nxt for makeing some kind of research token which is rewarded by participating in distributed computing projects.
An idea which I like even more, could be a nxt paralell blockchain, in which the coins are generated throw PoS and all the interests goes to a wallet controlled by the system which automatically rewards them to the "folders", this new paralell chain could be based in some distributed computing patafforms such BOINC, in which any researcher could submitt their workunits and they are processed by the network and even validated trhow some kind of decentraliced validation system "assuming majority of nodes areg ood guys, and the same workunit have been processed by many nodes, they should get symilar results, so if most results are symilar, we can say it's a good workunit"
Using this idea, we can use all unused cpu cycles by the time people is rewarded (Imagine that this tokens could become used to get disscounts in medicines discovered thanks to that or that big research companys decide to pay some fee to aceleratte their workunits processing)
This is a bit of my thoughts about this, it's only an idea, maybe I'm saying many stupid thinks, I'm not a coder like I said many times, what do you guys think? Could something like this work some day? I think too that new systems for distribution could be created, or POS concept could be modified to addapt to a better distribution model.
PD: Sorry if my english sounds strange and all the invented words
I didn't get the time to write it beautyfull or 100% correct, hope you can understand
