Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Supermikhail]

First, about my evolutionary idea. There are many very small mutations, and a lot of offspring (in the beginning, at least). I say evolution of animals went through all possible variations in the beginning before it got at least, I think, to worms. Or maybe even later. If there were isolated mutations (lagoon, lake) they didn't really influence the major evolutionary trends, and only rarely managed to become dead-end branches. All because a lot of very small mutations, but a lot of offspring to sort through them. And a LOT of time.

So, why do we still have all our creatures on the evolution to-do list? Why doesn't one of them become the most successful?

About sketching hundreds little things. With only as many creatures as we have I already feel inconsistency. Like they aren't connected. And it's kind of no fun without fleshing them out. Also, I could really use a planet where they live.

[Shoku]

Yeah, we're simply not going to through this with a world that's got reptiles that came from some worm while there's a whole group the size of mammals that came from a jellyfish. One thing is going to become more successful but right now we haven't fleshed out the ecosystem enough. Can you picture one of these things taking over what another one does right now? Probably if you pick two of the ones that just try to eat the same kind of thing but really what is there but the size of things they eat?
We need to address more of the prompts to get to a better point for that.

-

The simple sketch doesn't have to be the end point. You can draw the general shape of a dog in a few seconds but you could spend hours and hours drawing out the details of variations of dogs. The simple outline would quickly get us into new territory and then we can flesh it out from there. Stuff IS disconnected right now because like you've said, life tries everything. We need to get a little closer to that sort of everything before we can thin the herd and cause the more interesting developments to take place.

[Supermikhail]

So you're saying right now critters evolve without much pressure of natural selection? Then why do we get to Donutsaur and Octobrachius? These are big! You don't just come there. You've got to have hard selection!

... I know why. Because I am human. And I'm going to concentrate on one thing, and ignore other, maybe obvious, alternatives. So you want many small variants, and I just can't think of them.

Also, last night I asked myself what it is about this project that might attract other people to participate. Is it funny? Is it informational? Is it fascinating? Is it entertaining? And the way they are right now, our critters just aren't entertaining to draw and appreciate in the drawings (there's some interest in 3D visualisation, I believe).

[Shoku]

No, there's obviously selection. Right now there are enough bacteria of some of the more common species on Earth that basically every possible single base mutation happens on a fairly short timescale. (Can't recall if it's a few weeks or more like a year.) They literally try everything- but that doesn't actually change them a whole lot. You try everything in reach, natural selection takes a turn, you try everything in reach again, and so on. After a thousand generations your line ends up pretty different from the line that came from your twin sister (or not.)

But as for the size that's a matter of having organs. Bacteria colonies have physical limitations and I judged the size of ole burger particularly on how big the exposed ring would be compared to the unexposed area of cells. Plus some bias from C. elegans.

-

I've done enough art projects to know that you don't really have such extreme tunnel vision. If you're having a hard time thinking of other things go back to tactile paper and pencil and just draw like that. I made three really different forms without even adding any significant organs to octobrachius and they were different enough that I had to think for awhile. Changing just one or two things subtly is easier, is it not? I went with extreme because you seemed stuck not seeing anywhere else it could go but you've become almost hostile since then.
But hey, if you can't do small then don't. Big is fine and I've got all those prompts that basically ask for fairly big changes.

-

Fascination is the first draw but if you like hearing me talk about how life works it's informative too.
The lack of motivation to do this isn't a matter of things being un-fun to draw. The project has been slow and instead of making things go faster we just wait instead. You want things to happen without doing anything. I'm familiar with the feeling but I also know that nothing will happen if I don't keep trying.
I can't really fix this so if you want to be leave the project thanks for your contributions so far.

[Supermikhail]

Yeah. A break, I guess. In the past couple of days I was scared shitless of this ongoing tribulation, so I just didn't want to do anything. All mostly because I'm a useless sociopath; that's of no concern to this thread, but maybe with this disclaimer it'll be easier to understand my situation. So, today I went to the place of tribulation, and it's a little better than I feared it would be, so I come here. I'm not going to visit often, because I don't want to be involved deeply with something besides the tribulation. Also, I desperately want to learn programming, so that's another chunk of my free time.

By the way, you may remember of my talking of procedural generation in relation to this thread. I've dwelt on this idea some more, and decided it could work like this: a program takes input of a set of "genes"*, arranged in a coherent genotype, and "producing" a viable "phenotype". Then it goes through the "genes" and on a random factor applies "mutations" to them. Then it generates the "phenotype" for the mutation again, going through a set of tests to evaluate the viability of the "phenotype" on each level of generation. It then outputs the "phenotype" in some human-readable form, and at this point we are supposed to take the creature and decide if we like it and if it's going to be successful in struggle for survival, and we want to evolve it further.

*Everything in quotation marks means a mathematical abstraction with some simplification of the concept. Problems that need to be solved here include, mainly, how to generate a "phenotype" from a "genotype", what the human-readable form should be, how "genes" should be represented and what their mutations should be that a computer would be able to apply, and what the viability tests should be that the computer would perform during the generation of a "phenotype".

Why I thought such a program is possible? Because of DwarfFortress. Even if it'll melt your computer around the level of a worm.

Although today I've encountered a video that confused my thoughts. So all animals don't come from a single origin?


Some more on the topic of software. I recently discovered such thing as Google sites. Basically, a free website for a group project, even better than a blog or a shared document. There are templates, and I found the most useful to the project to be "Community Wiki" and "Project Work Site", and the like. Of course, only a bookmark until my life becomes more stable.

With this bouquet of ideas, cheers!

[Shoku]

Yeah, it was easy enough to recognize that you've got other things on your plate, particularly because you've either already said it or came close enough to saying it.

-

Converting a genotype into a phenotype is a monsterously enormous project. It's trivial to read a gene out and say what string of amino acids it will translate to. Slightly more bothersome to work out a few things that can modify that but still doable. What's completely undoable is figuring out what that protein does. You have to figure out which sections will be attracted and then bend it around so they're together in 3d space. This is still doable with computing power but not this century on a single personal computer.

http://fold.it/portal/info/science

Then when you have that you need to figure out how it interacts with other proteins and this is basically the point at which any efforts would utterly fall apart. We can't currently simulate neurons accurately even in programs that are keeping track of every atom in a system. We wouldn't be able to actually tell what the impacts were.

Now, you clearly weren't proposing such a lofty simulation of so much of a planet on atomic scale but how in the world can you abstract that for working out the shape of a body? It would be pretty easy to write a little program with some predefined parts and just randomly swap them out but that's already out there and it would stop you from ever getting anything but those predefined parts. How could you take some minimal fleck of the skin of a creature and turn it into a tooth or nail? How could just then say that the nail got thicker and thicker until it was a hoof? How could you make some piece of a body reverse the order of it's layers (that's roughly how we got corn) or take those earliest parts of an eye and decide to take it in the camera direction or do something more like the arthropods do? If you try and fit this process into a linear sequence of instructions it is wildly complex but for humans that can already mentally render 3d shapes while simultaneously thinking about the myriad of physical limitations (light goes straight, round structures can handle more force than harsh flat ones, etc) the real limit just becomes making yourself do it.

It's a lot of work to train just one person to think about every sort of process involved though and we can only create so many sorts of things. This is why I wanted this done by a group of people- you could get a lot of the creative efforts from many people and if it moved along quickly enough the group could even make the decisions about all of those lofty things we haven't even started on here.

-

Programming eh? Need to learn a particular language or just want to learn the process for now?

-

ANIMALS very much follow and come from the branching ancestry tree format. In the earliest days of unicellular life there was a lot of just using any DNA the cells got their hands on- it took awhile before there were any kinds of mechanisms to prevent that and even once there were they'd have had to make their way into other cells and so forth. Eukaryotes have this obvious little thing going on in their cells: organelles. Some of these might have come from folding a lot of the membrane inward and then pinching it off so as to have different compartments inside of the cell but at least a few are very clearly originated from other cells, mitochondria being the most frequently used example but things like eukaryote flagella (very very unlike bacterial flagella) being somewhat likely candidates. Many of these structures can be grown from scratch right off of the genes in a cell's chromosomes but why would single cells want to arrange their DNA in all of these different chunks? Well if you were merging whole cells you'd end up with two different strings of DNA in the combined cell. Maybe you'd want to fuse them together or maybe not but if they were loops like in bacteria there would be no exposed end to weld onto another.

But basically what that's saying is that while life was still really working out what a cell IS and getting the basics going the chemicals ended up all over the place. Now, just like you can understand the difference between your great*5000 grandpa not being the same thing as the chimpanzees around today you can get that the "bacteria" way back then aren't the bacteria of today, or even a couple billions of years back. Everything keeps evolving and while the bacteria pretty much kept within the niches we know bacteria exist in they did keep evolving their cellular machinery.

It's simplified, watered down, and at some point downright stupid to think that there was some murky water that went through all of the chemistry to produce a bacteria almost exactly like modern ones and that it then ceased entirely as the cell bloomed into such enormous numbers as to eat any other potential cell. There is a point where that happened but there was also a transitional sort of period where things were starting to form stable membranes and such but where the DNA from one could just as easily end up in another.

But the distinction of animals I made earlier is pretty easy to follow if you know what you're looking at. Where that video mentioned opisthokonta you're got a few protists, all fungi, and animals using these rear propelled sperm. Seeing as no fungi or animals can ever step back through the tree of evolution to become a protist (you could mix those three up in any order in that sentence and have it technically work but you know what I mean,) they're clearly dealing with the branching lineages of the tidy ancestral descent we're used to at that point. So yes, animals have a single ancestor species, and depending on how the numbers actually worked out perhaps a single shared ancestor, but not at the very base of the tree of life. If you want to try and trace the path back you first run into all of those problems of cells taking up other cells and then even further back run into problems of all of that even less rigid DNA transfer. There are obviously many points before sexual reproduction where you can narrow things down and say that this individual or that had some particular mutation that our heritage didn't get from any place else so they would be a single ancestor of ours but a universal ancestor? Well, how could you even tell a cell apart from the things that lead to it at that point?

-

I may look into and set that up. I never said I wouldn't do these things, just that I wasn't feeling up to it at the time.

[Supermikhail]

Hey, it has just occurred to me to look up "Gene" in Wikipedia, and to my pleasant surprise, "Following the sequencing of the human genome and other genomes, it has been found that rather few genes (~20 000 in human, mouse and fly, ~13 000 in roundworm, >46 000 in rice) encode all the proteins in an organism." Relative to modern computer processing power, this, as just a number, is quite nice.

My ignorant, and daring vision here is such: scientists say that some gene is responsible for eye colour. That is, it appears that you can connect a segment of DNA to such a macroscopic trait as eye colour. I realize that developmentally that's not so easy. Still, I insist on viability of such a project. Cool guy Carl Sagan says that in theory human brain wouldn't be able to comprehend even a grain of salt in all its microscopic complexity, because there are fewer information-holding elements in the brain than atoms in a grain of salt. But we don't need to store the properties of all the atoms, because they are governed by a few physical laws based on which we can predict properties of a single atom, if we want to, or the grain as a whole... I guess, I didn't really need to say all that. Well, let that be a necessary preface. What I mean is that the program will have to make compromises, go through several levels of abstraction and at some points be completely incomprehensible. Also, go through endless loops to produce planet-sized cancer blobs (of course, loops should be breakable then) and have untraceable bugs translating perfectly into bizarre organic structures.

Some of... well, I'm going to call it an abstraction, for the lack of a better word... some of abstractions could include taking a cell as a single unit and ignoring all of its inner structure besides the DNA. Or starting with an even bigger structure... because the evolution of a cell into a bicellular organism turns out to be the quantum physics to the Neutonian, if we were to apply physics analogy to our biological laws.

In any case, after I have read the Invertebrate Zoology, I should probably get to embryology or some kind of genetics.

Now, about the actual programming. First, my reply today probably gives away a pretty cheerful mood I'm in, brought about by the guy supervising my graduate paper giving a "go ahead" for the chemistry game I'd like to program. And I'd like to program it in C++ because I've studied it intermittently, although I just couldn't handle advanced inheritance when I got to it in my textbook. Of course, I'd be happy to hear your suggestions on the programming language, and learn it, but only when I get into a more stable phase of my life. And, I guess, the development and even serious research for the mutation program isn't going to start soon.

-

Well, "at the time" can be different things.

[alfie275]

By chemsitry game do you mean something like http://www.spacechemthegame.com/ ?

[Supermikhail]

Edit: Crap. Didn't watch the trailer. Yeah, something like that, but worse.

Anyway, Shoku, would you say it's possible to induced-dream something for a thread like An Otherworldly Ark?

[Shoku]

While that is the number of genes in an individual we have a great deal more DNA than that and many more proteins present. In the large stretches of DNA between the parts that become protein you have numerous regions that ultimately influence how much of that protein gets made (either many times more than it would without their aid or as much as complete deactivation of the gene) and then within the gene itself you need to worry about exons and introns (translation: different parts get cut out in different cell types, so one gene can make a dozen variations of some protein.) and finally there is again that matter of the proteins actually interacting with various chemicals, which include other proteins.

-

We can't ignore the atoms because we need them all there to figure out what a protein does in the first place. If we were just mixing and matching existing and well documented genes then we could easily abstract out that level of detail but with development that are so many cases where how fast the group sum of some protein causes a chemical reaction matters that we absolutely cannot ignore the molecules.
Ex. All five of the fingers on your hand work out their location based on one particular protein. Others are involved in the process of course but this one particular cluster of cells makes the protein and then it leaks out of them into the surrounding tissue. At the highest concentration you establish where your thumb will be and at the lowest where your pinky will be. Now, how do you suspect we could model the loss of an entire finger during development based on mutations to that particular gene? The mutations could cause it to spread slower or have a different magnitude of activity with the target proteins it interacts with or even for new regions upon it to bind to some other molecule which prevents it from accessing the usual target. And the binding may not be a permanent condition so when they come apart some time later it could then go back and interact with the molecules for finger development.

After you've seen how it behaves with types of molecules and worked out various attributes of the interaction you can just run to equations for dealing with that but again, if we simulate mutation we would be creating molecules we don't really know anything about and so we would absolutely have to work out all of those attributes ourselves and seriously, the big life career researcher guys still spend years or at least months trying to do those sorts of things for just one particular protein.

Basically right now we've only figured out how to read a genome and we do a little bit of copy and paste between them to figure out other intricacies of how things work. We will have people writing genes from scratch long before it would be remotely practical for hobbyists to do. One day sure, but that day is still years down the road.

-

Oh, I would have just given an introduction to programming if you didn't know anything about it. Sort of a second hand experience of a class I took. Obviously you're beyond it.

-

I have no idea what you are asking there.

[Supermikhail]

Okay. How about not real DNA? Not real proteins. Not real creatures. I, over the course of this discussion, have somehow become more interested in abstractions and developing a macroscopic construct based on a microscopic bluebrint. There's got to be a way for that. Say. The program gets a blueprint. First, it goes over it and gets all the active zones. Then from these zones it reads a list of building blocks which... Hold on. That game, Foldit. There's those two buttons, and if you press it the program will fold the protein on its own, it seems to me. Although, I didn't get far in it... Anyway, for the purposes of mental health, the blocks come with more geometric constraints than real aminoacids. With such hard constraints, that even a computer can fold a tertiary structure... Hm. Then, I guess... for simplicity's sake it goes through all the tertiary structures and first tries to fit them to one another, then to a predefined number of the abstract cell's chemical constituents. Although, at that point, not only geometry, but reactivity, too, should be taken into account. So far, the success of the program depends on the power of the 3D engine, the comprehensiveness of the database of reactivity, and how big abstract proteins are allowed to get. I guess it would be pretty important to check if abstract proteins can interact with abstract lipids of the abstract cell membrane and the abstract DNA... Ah, my head is hurting already from all this speculation.

Anyway, induced dreams. "Inception". And Carl Sagan. I just thought it'd be cool if I could persuade my subconscious to work on An Otherworldly Ark while I sleep. Unfotunately, this having any effect on the reality requires being able to recollect one's dreams, and I currently am having trouble with that. A few dreams that I can remember from the previous year, I guess, featured me as a protagonist of some weird computer games. But to bring more imagination here would be beneficial probably. I tried it and imagined for some reason a stretch of sea studded with donutsaurs whose texture was like of a jellyfish, and for some reason they were very white but transparent underwater, I think that means they reflected light very well, at the same time having high Fresnel. I'm currently trying to replicate the effect, but Blender doesn't want to cooperate as usual.

[Shoku]

I went over that though. How can you grow some fleck of skin into a tooth and then into a fang? Living things do this via shifts in protein trascription and and activity. They make some duplicate of things they already use or put them to use in different tissues with all of these different things in a cascade of activity. I can't think of how else you could define a body for a program while allowing the parts to have directional development. You can play mr potato head easily enough but people already do that with actual symbolism behind the result they pick so why bother?

No, you didn't look into foldit enough. They take actual stretches of DNA that they've read but don't know enough about. They couldn't get powerful enough super computers to handle all the protein folding they wanted to compute so they send it out into the cloud and have countless home computers use idle cycles to do the computing. And that right there? That's a full group project from no doubt very devoted students. It takes more than a bachelor's degree in chemistry to know about all of the nuances they had to program into that and you definitely need some pretty experienced programmer to make something that will do such heavy computing on so many different hardware configurations.

But yeah. It would be possible to make programs do this. What I'm saying is that it is entirely unfeasible for the two of us, or even if there were a dozen others involved. I didn't set out to create a project that takes experts to execute, just one that takes interested people with a speck or two of creativity.

-

But to what end? It might make for a nice scene but I don't see how it would help us with diversity, which is what we're really lacking for a good scene of early life in this project.
Are you still actively thinking about new forms? Are you stuck not knowing how to do something?

[Supermikhail]

Yeah... That's... a conundrum. I just thought how useful it would be to have such an abstract program like I've described. You know, just in case we come into contact with an extraterrestrial form of sentient life. As far as my understanding currently goes, there's really few paths life can take from a certain starting point, so theoretically, we could know something useful about the extraterrestrial form having only just met them, based only on the geological history of their home planet. What we can't do here, because this project isn't very scientific as we can't account for everything. Namely, I am concerned about the burger. I don't think it should exist, with what I know now. It's supposed to be a colony of cells that for some reason go in a shell. And it's supposed to have jets. Jets require a lot of tissue specialisation, and living in a shell like that requires a good system of nutrient transportation, because most cells are just cut off from any nutrients. I believe you raised at least one of these points when I first came up with the burger. Well, these things have finally made their way into my brain, and I can't live with them. I'd really like to look deeper into them and possibly start from a scratch.

[Shoku]

Not so much I'd say. Most of the animal life on Earth has all of it's sensory organs on the head because of that whole worm origin thing. Worms have long bodies with the mouth in front and move in that direction and that early design decision has stuck around to the extent that basically everything has a "face." And just think about all of those mass extinctions- if instead of mammals taking over after dinosaurs instead birds moved into most of the niches first. Mammals might have just stuck around as some small breastfeeding group and never even switched away from laying eggs. Or how about if instead of fish developing for land it was lobsters that first adapted for life out of the water. How about if life wasn't worm based at all? Or most strikingly if the metazoans had first appeared in a strikingly different order.

Every time life does one thing it cuts off such a huge number of other things. I used fairly familiar examples in all of that but think about how animals like the duck billed platypus were actually pretty typical of early mammals and you should be able to see how much potential there is in obscure groups developing wildly into new forms. There's so much life on this planet that we're certainly not limited to a small number of options but rather confronted with an open sea of them so vast they seem infinite.

-

But yeah, if you want to give it another go after the impact of listening to Sagan we could. The origin of shells was actually one of those points where one sort of creature started to dominate the rest so we could leave that out of the mix for awhile and for movement we could stick with more of the flapping sorts of motions. Without the shells you won't need much aid to buoyancy anyway so things can move slowly at first and then later take to faster forms of movement.

Don't be too afraid to adapt aspects of what we've already done here. The forms don't need to be especially unlike everything else and hopefully without shells and jets I'll feel comfortable contributing at a similar pace to you~

[Supermikhail]

Hm. Now I'm having difficulty thinking of something different now that I know how it went on Earth. I'm stuck at thinking that for our ordinary cells there could be kind of two ways - a sphere opening upwards, and a sphere opening downwards, like in the case of your bulbousaur.

However, burger overhaul got me moving, it seems. At a colony level you might not need a developed transportation system, it seems. There's an in-cell transportation system in some protists, I think. At least in a... four layer colony outside cells could transport nutrients to inner cells with some success. But if the colony gets bigger, they have to get and transport nutrients very efficiently. So, suppose the... trisgea shelled species finds itself a symbiont - regular, not shelled, cells. The latter form an outer layer around a reinforced core, and provide it with nutrients. In turn, trisgea provide them outer cells with protection of sorts. If something attacks the colony, it can strip it of a softer outer layer but won't get into the protected inner layer, where there are soft cell "caches" which repopulate the outer layer after the attack.

If this design passes your critique, it can develop several ways, for example "ring", "stripe", "spiked wheel", "dodecahedron"... come think of it, a lot of geometric structures with hard edges. Although, turning one of these colonies into a whole organism is still a problem.