Will we run out of chemical space?

[mississippichem]

Many people are not familiar with the concept of chemical space. In short, chemical space is the set of all "possible" compounds. The chemical space is quite large. For example, there are about 10²⁹ stable derivatives of n-hexane with 150 substituents or less (1). But one has to consider that the space of compounds that can actually be made by some real synthesis must be much more limited than that. Even further, the number of actual distinct chemical topologies must be even smaller. By topology I mean distinct functionalities that are not redundant. For example, just building longer and longer alkanes is not interesting.

 

My question for the thread is, will organic chemists run out of interesting chemical space to explore? Ever? Perhaps in the distant future?

 

The second part of the question I'll propose is whether or not all the easily accessible functional groups have been found. There will be a degree of subjectivity in anyone's definition of "easily accessible" but I think this can make for an interesting discussion anyway.

 

Your thoughts?

 

 

 

(1) Christopher Lipinski; Andrew Hopkins

Nature 432, 855-861 2004

[A Tripolation]

Is your conversation/thoughts in any way analogous to this?

Edited March 2, 2012 by A Tripolation

[elementcollector1]

Many people are not familiar with the concept of chemical space. In short, chemical space is the set of all "possible" compounds. The chemical space is quite large. For example, there are about 10²⁹ stable derivatives of n-hexane with 150 substituents or less (1). But one has to consider that the space of compounds that can actually be made by some real synthesis must be much more limited than that. Even further, the number of actual distinct chemical topologies must be even smaller. By topology I mean distinct functionalities that are not redundant. For example, just building longer and longer alkanes is not interesting.

 

My question for the thread is, will organic chemists run out of interesting chemical space to explore? Ever? Perhaps in the distant future?

 

The second part of the question I'll propose is whether or not all the easily accessible functional groups have been found. There will be a degree of subjectivity in anyone's definition of "easily accessible" but I think this can make for an interesting discussion anyway.

 

Your thoughts?

 

 

 

(1) Christopher Lipinski; Andrew Hopkins

Nature 432, 855-861 2004

 

Well, I don't know. A vision I have for the future of chemical synthesis is a sort of super-optical tweezers, whereby individual compounds can be taken apart molecule by molecule (and sorted, by nanobots). Then, any chemical is synthesizable.

Of course, this is probably a pipe dream. Then again, wasn't everything in science like that before someone figured it out?

 

 

[mississippichem]

Is your conversation/thoughts in any way analogous to this?

 

There is some similarity. Here the limit is the number of realizable bond configurations. It is my opinion that there is a fundamental limit to the number of types of compounds. That number is just incredibly huge.

 

Another more subjective part of what I'm asking is, will it continue to be interesting? Which makes what I'm talking about different than what you mentioned in a way.

 

Well, I don't know. A vision I have for the future of chemical synthesis is a sort of super-optical tweezers, whereby individual compounds can be taken apart molecule by molecule (and sorted, by nanobots). Then, any chemical is synthesizable.

Of course, this is probably a pipe dream. Then again, wasn't everything in science like that before someone figured it out?

 

Atom by atom synthesis may be possible one day in the distant future. Whatever those methods are their limitations are likely to be completely different than the ones we have now. That's an interesting way to look at it though, from the standpoint that future synthetic limitations will be less. But we still have to consider stability. Compounds that have a short (unusable) half-life now will still be just as short lived no matter how finesse our synthetic methods. Will we run out of novel compounds even in the event of ever-better synthetic tools?

[JohnStu]

Advancement in atom by atom synthesis would be something nano technology could benefit, right. The cons I see with atom by atom synthesis is that they need to be done in mass production - having thousands of same chemicals alinged all together attached with the same new atom at the same spot, and then the next batch of chemicals come to be done the samething. Then this method could excel the synthesis methods they currently widely use. The advantage of atom by atom synthesis over regular synthesis methods would be it could be really precise and flawless in the near future. It could also allow creation of atomic processors perhaps someday.

 

 

[mississippichem]

Advancement in atom by atom synthesis would be something nano technology could benefit, right. The cons I see with atom by atom synthesis is that they need to be done in mass production - having thousands of same chemicals alinged all together attached with the same new atom at the same spot, and then the next batch of chemicals come to be done the samething.

 

Who knows how such a method might work? We are so far away from anything even similar that any guess as to the engineering that might go into it is pure speculation.

 

Then this method could excel the synthesis methods they currently widely use.

 

Not sure I understand. Atom by atom synthesis wouldn't involve hardly any of the methods we currently use. This sentence doesn't follow your previous ones.

 

The advantage of atom by atom synthesis over regular synthesis methods would be it could be really precise and flawless in the near future.

 

No way. We are nowhere near atom by atom synthesis and we don't really know if anything like that will ever be achievable or even desirable/practical. You're really speculating quite intensely without much support. Also, the thread is about the chemical space in particular, though some speculation about future synthetic methods may be relevant.

Edited March 2, 2012 by mississippichem

[StringJunky]

Well, I don't know. A vision I have for the future of chemical synthesis is a sort of super-optical tweezers, whereby individual compounds can be taken apart molecule by molecule (and sorted, by nanobots). Then, any chemical is synthesizable.

Of course, this is probably a pipe dream. Then again, wasn't everything in science like that before someone figured it out?

 

You made me think (a rare thing!). Do chemists have "cyber laboratories" now where they can virtually test prior unknown combinations of elements and compounds under specified conditions and see the resultant products?

 

I envisage a chemical database with every known necessary parameter for each element and then being able to chuck two chemicals into a virtual test tube and the software, once all conditions have been inputted, tells you the result.

 

How far down the road is chemistry with that scenario now?

Edited March 2, 2012 by StringJunky

[John Cuthber]

 

No way. We are nowhere near atom by atom synthesis and we don't really know if anything like that will ever be achievable or even desirable/practical. You're really speculating quite intensely without much support. Also, the thread is about the chemical space in particular, though some speculation about future synthetic methods may be relevant.

I'm not so sure about that.

http://www.almaden.ibm.com/vis/stm/atomo.html

The issue is that making just one molecule of something isn't going to achieve much.

[Xittenn]

You made me think (a rare thing!). Do chemists have "cyber laboratories" now where they can virtually test prior unknown combinations of elements and compounds under specified conditions and see the resultant products?

 

I envisage a chemical database with every known necessary parameter for each element and then being able to chuck two chemicals into a virtual test tube and the software, once all conditions have been inputted, tells you the result.

 

How far down the road is chemistry with that scenario now?

 

Rosetta Commons is my favorite current example of this--that I still can't use. It's not as you say but it folds proteins and performs binding. EPM solvers and so forth exist, and my engine that I've been working on for two years now will--maybe in ten give or take five--one day be operational. The answer is basically now but has been in infancy for some time . . . . . .

 

I wonder how long it will take the first chemical space cyber run to complete!

[CaptainPanic]

There is some similarity. Here the limit is the number of realizable bond configurations. It is my opinion that there is a fundamental limit to the number of types of compounds. That number is just incredibly huge.

 

Another more subjective part of what I'm asking is, will it continue to be interesting? Which makes what I'm talking about different than what you mentioned in a way.

You cannot get bored. If you organic chemists are done playing with your ideal reactions that start with purified reactants, you can finally get started on the world.

 

Here's a picture of what the real world looks like:

 

It looks so simple, but it is not.

 

A plant contains the entire periodic table, and probably also all types of bonds that are every written down in an organic chemistry book. Plant chemistry is so un-boring that I would even call it exciting.

 

What you get then is your 10²⁹ components all in 1 reactor, doing some funky reactions to each other. Logically that should give you 10²⁹²⁹ possible reactions to explore. No new chemicals maybe, but plenty of action.

[hypervalent_iodine]

You made me think (a rare thing!). Do chemists have "cyber laboratories" now where they can virtually test prior unknown combinations of elements and compounds under specified conditions and see the resultant products?

 

I envisage a chemical database with every known necessary parameter for each element and then being able to chuck two chemicals into a virtual test tube and the software, once all conditions have been inputted, tells you the result.

 

How far down the road is chemistry with that scenario now?

 

Yes, but they're not perfect. There are a number of groups that use computational models to design certain compounds for use as high energy materials, drugs, etc. It's used extensively in protein docking studies as it helps in designing substrates that will bind to the active site of various proteins. There is also combinatorial chemistry, of which there are mixed opinions depending on who you ask.

 

A major and related problem as I see it is that current research is being forced to focus on a very small subsection of chemical space. A large majority of new compounds are designed with the purpose of drug applications in mind and these are required fit into a very restricted window of parameters pertaining to things such as molecular weight, lipophilicity, etc. Funnily enough, a large number of drugs that were designed, marketed and successful prior to this being implemented are so far outside of these limits that if you were to suggest their use today as a new compound, you would probably be laughed at. The problem with this is that it narrows the focus of chemical space to the point of ignorance and this can be particularly detrimental to the application side of things. Eventually, research will have to start venturing outside of this as screening methods become more thorough and more efficient. In fact, even now I have seen a few papers where this issue has been realised. It will likely take a little more time before anything is properly done about it, however.

 

You cannot get bored. If you organic chemists are done playing with your ideal reactions that start with purified reactants, you can finally get started on the world.

 

Here's a picture of what the real world looks like:

 

It looks so simple, but it is not.

 

A plant contains the entire periodic table, and probably also all types of bonds that are every written down in an organic chemistry book. Plant chemistry is so un-boring that I would even call it exciting.

 

What you get then is your 10²⁹ components all in 1 reactor, doing some funky reactions to each other. Logically that should give you 10²⁹²⁹ possible reactions to explore. No new chemicals maybe, but plenty of action.

 

Entire periodic table is a biiiiiig stretch there. In fact, it's plain wrong. It contains most of the useful elements to an organic chemist, but by no means does it contain all elements, ever.

 

As for the potential for developing reactions based off of what occurs in plants. Certainly, exploring the mechanistic aspects of these reactions is useful in the sense that it opens the way for developing new reactions ex vivo. Actually utilising these reactions can be problematic. While bio-mimetic synthesis is a great idea in theory, in practice it is often superseded by more convenient and preexisting synthetic approaches. Quite often they simply don't work at all because trying to recreate the reaction conditions can be quite difficult to do. You could potentially exploit the plant or at least the enzymes to do the reactions for you, and the latter is certainly an active area of research, however they both come at a cost since enzymes are expensive and separating compounds from plant material is both time consuming and wasteful.

 

No new compounds? Quite the opposite! There is a large area of research focused on structural elucidation of secondary metabolites from plants, etc. It's been going on for some time too, so there is also a lot of research out there. Some of my favourites are the hetisine type alklaoids. They have this wonderful heptacyclic fused ring system that is an absolute nightmare to draw and even worse to make. They've been known since the 50's and since then, over 100 have been discovered. You'd think we'd have making them down to a tee, but in fact there has only ever been one made successfully and it happens to be the least structurally demanding of the lot. No points for guessing that these are part of the focus of my research. Another very interesting avenue is marine sponges. There is a group a level down from where I work that uses these to extract all sorts of whacky compounds from.

[Appolinaria]

A plant contains the entire periodic table

 

 

Definitely not, haha.

[CaptainPanic]

Ok, I admit that I exaggerated. A bit. A lot. I figured that the numbers were so big, that it wouldn't matter if I was a few orders of magnitude off.

 

Still, it's quite common to measure up to 30 different elements in plant material, and that's for untreated wood from unpolluted land. If you get it from the real world, and you add pollution and you obtain the material from some secondary source (waste wood or something), it gets a more complicated. Also, if you buy wood or something in bulk, it also contains soil (rocks and sand).

 

Still, some elements are so rare and have such a short half-life that they are hardly ever found. And I also never heard of noble gases in plants. My statement was obviously wrong anyway. I never meant to be serious.

 

Regarding the reactions: I was talking both about the biochemistry (enzymatic reactions etc), and also about the reactions that you can do in chemical reactors, possibly at higher temperature and with some catalysts, using plant material as reactant.

 

[edited because I wasn't happy when I read my post again]

Edited March 2, 2012 by CaptainPanic

[John Cuthber]

Captain,

you might want to think about how "organic" chemistry got it's name before telling us to look at plants.

Everyon else

If you didn't find, for example, uranium and gold in that sample, it's just because you didn't look hard enough.

I'm prepared to accept that the "practically non existent" elements like At and Pm might be absent

[mississippichem]

Yes, but they're not perfect. There are a number of groups that use computational models to design certain compounds for use as high energy materials, drugs, etc. It's used extensively in protein docking studies as it helps in designing substrates that will bind to the active site of various proteins. There is also combinatorial chemistry, of which there are mixed opinions depending on who you ask.

 

I tend to be of the combinatorial pessimist crowd. It is my opinion that in order to do combi-chem effectively, your libraries would have to be orders of magnitude bigger than anything anyone has attempted to date. I had a blog entry about this. I never blog really but here it is none the less, The d-blo[g]ck. There you will find my brief synopsis and a link to here were the issue is detailed. In short, the problem is that combinatorial libraries compose such a small portion of the chemical space as to be ineffective. So in essence the process is like trying to map the earth with a microscope, i.e. you can't see enough of the whole at one time to draw any useful information.

 

Now as far as computation goes I am quite the optimist. Of course I have a pretty major bias as I'm involved with computational work . I think the approach of finding a structure that fits a set of predetermined properties is much more realistic than making a load of compounds and hoping that some of them have said properties.

 

A major and related problem as I see it is that current research is being forced to focus on a very small subsection of chemical space. A large majority of new compounds are designed with the purpose of drug applications in mind and these are required fit into a very restricted window of parameters pertaining to things such as molecular weight, lipophilicity, etc. Funnily enough, a large number of drugs that were designed, marketed and successful prior to this being implemented are so far outside of these limits that if you were to suggest their use today as a new compound, you would probably be laughed at. The problem with this is that it narrows the focus of chemical space to the point of ignorance and this can be particularly detrimental to the application side of things. Eventually, research will have to start venturing outside of this as screening methods become more thorough and more efficient. In fact, even now I have seen a few papers where this issue has been realised. It will likely take a little more time before anything is properly done about it, however.

 

It always makes me laugh when you read a great synthesis paper, especially natural products synthesis and the author mentions that this compound might find application in the pharmaceutical industry. Anyone who hopes to be able to scale up a 22-step synthesis producing a compound with five chiral centers and three bridged rings is fooling themselves, given the current state of process chemistry that is.

 

I think that part of the reason that so much of the chemical space is unexplored is the way we do chemistry. No one does a novel synthesis without some form of literature precedent anymore. This is understandable as people do not wish to take risks with their livelihood.

 

Take this compound for example:

 

 

Bad article with an interesting compound

 

I absolutely abhor the article and what it's about with respect to education but ignore that and look at the structure. This compound is simple right? The symmetry is high. The carbon backbone is not complex. This an example of those inaccessible areas of the chemical space. I can almost guarantee that no one will be able to find a viable synthetic route to it.

 

I'm not so sure about that.

http://www.almaden.i.../stm/atomo.html

The issue is that making just one molecule of something isn't going to achieve much.

 

That is quite interesting but are those not just single atoms and diatomics intercalated onto a metal surface? I see how this type of technology could eventually lead itself to something that resembles atom by atom synthesis though. Nice link, thanks.

[Xittenn]

Given the present state of calculating real time electrodynamics I would imagine it would be over a hundred years to complete a run over chemical space--most likely much, much more. We could start 1% of the worlds computing power on this now and probably not see the end until 3000! I mean look at proteins it isn't simply constitution but it is also all of the folding and the slight variations. Make that 4000 unless serious advances are made in computing. I don't even know how to put an upper bound on this?????

 

What compound has the highest number of atoms in it?

[John Cuthber]

A single crystal of silicon, as made by the electronics industry, is a single molecule and has an awful lot of atoms. The same goes for diamonds and at least some other crystals.

 

It's not just a matter of how many atoms are present in a molecule, but how much variety.

There's only 1 isomer for propane, but there are two different chloropropanes and rather a lot of bromochloropropanes.

 

Does anyone know of a general expression for the number of alkane isomers as a function of number of carbons?

[Xittenn]

I think of crystals as repeating decimals. As for the general formula I suspect there would be a couple that could be developed and would most likely consist of the most appropriate arrangement of nCr as in combinatorics; I've personally never touched combinatoric chemistry. Do you know the answer or was the question genuine? : D

[John Cuthber]

I think this

http://en.wikipedia.org/wiki/Combinatorial_enumeration

is the right branch of maths. Unfortunately I don't understand a word of it.

So, yes, it's a genuine question.

[Xittenn]

Try these two and reference a text to fill in the blanks:

 

http://actachemscand.org/pdf/acta_vol_04_p1450-1463.pdf

 

Lygeros, Nik, Paul-Valère Marchand, and Marc Massot "Enumeration and 3D representation of the stereo-isomers of alkane molecules." Journal of Symbolic Computation 40.4–5 (2005): 1225–41. Print.

 

 

The second does a pretty good job of detailing the history.

 

**but skips derivation up to more current stereoisomer research . .. .

 

***when did Alkyne<=>Alkine :/

[John Cuthber]

Thanks for that.

[mississippichem]

Given the present state of calculating real time electrodynamics I would imagine it would be over a hundred years to complete a run over chemical space--most likely much, much more. We could start 1% of the worlds computing power on this now and probably not see the end until 3000! I mean look at proteins it isn't simply constitution but it is also all of the folding and the slight variations. Make that 4000 unless serious advances are made in computing. I don't even know how to put an upper bound on this?????

 

What compound has the highest number of atoms in it?

 

Well in the pure consideration of a single molecule in a vacuum I don't think there is an upper bound to molecular weight. However there are real synthetic limitations.

 

Take for example a simple linear alkane. In theory, though it's not too practical, we can synthesize alkanes of considerable length from methane by a series of radical reactions; each successive reaction having a radical termination step that yields an alkane longer than the last reaction (I know the lengths will really be more of a statistical distribution but we'll not consider that). Eventually we get to the point where we have to apply a lot of energy just to keep the alkanes in the gas phase during the reaction (somewhere around C₂₀H₄₄) so we switch over to solution phase, bring in a Zielger or metallocene type catalyst and starting making polyethylene which is just really high molecular weight alkanes, up to a few million g/mol. Even here we still run into a practical limit though as at some point we don't have a high enough molecular weight liquid solvent to solvate an ever growing polyethylene chain.

 

What I'm getting at is that there is no theoretical limit to the molecular weight of a compound but there are very real practical limits.

 

Try these two and reference a text to fill in the blanks:

 

http://actachemscand..._p1450-1463.pdf

 

Lygeros, Nik, Paul-Valère Marchand, and Marc Massot "Enumeration and 3D representation of the stereo-isomers of alkane molecules." Journal of Symbolic Computation 40.4–5 (2005): 1225–41. Print.

 

 

The second does a pretty good job of detailing the history.

 

Interesting, +1.

[CaptainPanic]

Captain,

you might want to think about how "organic" chemistry got it's name before telling us to look at plants.

I know that - but good to point it out again.

 

I think that at some point in the previous century we should have renamed it to "fossil oil chemistry". Plant material is not often used as a source for chemicals - especially not at large scale applications.

 

Plant material is obviously still used a lot - food, wood, paper, to name a few - but often it is left largely intact.