Skip to main content
Menu

Chemical News

A Retrosynthesis Contest

Here’s a retrosynthesis challenge from Merck KGaA in Darmstadt. They’re celebrating the company’s 350th anniversary, and this is apparently part of the festivities. Anyone can enter for free, and the company will choose up to 12 entrants to take part in the competition itself. As I understand it, each selected person/group will then be furnished with the structure and will have 96 hours to submit a proposal. These routes will then be tested at the bench in a contract synthesis lab. The winner (based on number of steps, overall yield, and isolated purity of the final product) will get 10,000 euros, which is not a bad reward.

There are, of course, crowdsourcing platforms for scientific problems, with Innocentive being the best-known. This challenge, though, doesn’t seem to be as much of a direct “How do we make this molecule” question (for example, they aren’t going to commercialize the compound or anything arising out of the challenge). The Merck folks tell me that one motivation comes from their recent purchase of Chematica. They would like to know what the differences are between human ingenuity in retrosynthesis and the software challengers to it, so they’re throwing this one open to the chemical community at large.

To me, this sort of competition is two competitions mixed together. The first is paper synthesis – how many reasonable route are there to a given structure? How many rough classes do these approaches fall into, and how many weirdo singleton routes show up? What sorts of reactions and conditions get proposed more than others, and is this some reflection of the “natural order of things” or just an artifact of how we all learn organic chemistry?

The second competition is reduction to practice, and that’s where things get randomized a bit. As one or two readers may have had occasion to notice, not all good paper routes work out exactly as drawn. Where, then, are the skidmarks left when the rubber hit the road – what reasonable steps turned out to fail, and why? Could these failures have been anticipated, or are they more cases of “Well, whaddya know”? Can reactions and whole synthetic strategies be evaluated for robustness before you even start, or not (yet)?

All of these questions, at both stages of the challenge, are of great interest for AI development of retrosynthesis, so I can see why the Merck/Chematica team would want to see what happens. (I would guess that the Innocentive folks have a good amount of data on these topics as well, although I don’t know what use is being made of it). I very much hope that we get full reports along the way from Darmstadt, and I encourage the company to keep everyone informed. An open database of proposals would be great, along with updates about how the actual bench chemistry performs as it happens. Good look to all entrants!

31 comments on “A Retrosynthesis Contest”

  1. Uncle Al says:

    These 12 candidates will be given the structure of a small molecule and have 96 hours to submit their proposed synthesis route” Cute, but not a study. Run the dozen, then open it to the world and get statistics on human realization.

    Suppose an undergrad says, “isn’t that a multi-component isocyanide reaction?” and it is. Does HR toss away serendipity (or genius) because LaDavid Ugi is formally unqualified?

    1. BGR says:

      I hope not. I’m an undergrad who used to work next to a group that specializes in isonitrile chemistry. Walked past a poster detailing every extant multicomponent isocyanide reaction every day.

  2. Anon says:

    As if synthesis was the rate limiting step to drug discovery and Pharma R&D productivity in general…

    Still, an interesting academic exercise.

  3. Anonymous says:

    As I recall, InnoCentive was created by Eli Lilly sometime in the 1990s and it began with synthesis problems. I signed up early on. At some point, it was spun out as more of a consortium with problems from numerous sponsors.

    I was always bothered by the terms: all IP belongs to the Sponsor, not to the Solver. Most of the awards were kind of meager, considering the cost of coming up with solutions. Some Problems required that the solution be reduced to practice in the lab and not merely a proposed solution; the expectation being that someone else would pay for all of the R&D (winning AND losing R&D) and then the sponsor would reap the benefit. Some of the early winners were from Eastern Europe; in one case, a group pulled the winning solution “off their shelf”; it was a compound that had been synthesized decades earlier (under Communist rule) so all they had to do was make some more and send it in – all paid for by the (Polish? Hungarian?) government.

    Even proposed solutions (on paper only) require time and effort and there are COSTS associated with searching the literature. At a university, “someone else” is paying those costs (usually gov grants). A private organization needs access to the literature to compete and lit access is getting more and more expensive.

    I think InnoCentive is probably a great deal for the Sponsors but I see it as another way to keep chemists (and other scientists) unemployed, scraping at the chance to get a $5000 payday.

    If things have changed, I hope that someone will correct my impressions.

    1. John Wayne says:

      I agree with you on all points. The terms and financial rewards offered by this idea were fairly insulting to the average chemist. My concern was the only folks who don’t understand the value of their own time would participate, diminishing the apparent impact of chemists to those running the show.

    2. Nick K says:

      I too think the rewards offered by Innocentive were insultingly small, especially for the synthesis projects requiring the submission of a sample.

    3. Isidore says:

      If you thought the rewards were small for your effort then obviously you would choose not to participate. For other chemists the rewards were, apparently, adequate to good so they participated. The difference, I suppose, between making, say, $100K/year vs. $15k/year. For the former $5,000 might very well not be worth the time and effort, for the latter it might.

  4. anon2 says:

    Paper routes are all fine and good but all bench chemists are not even remotely created equal. Give a crappy chemist a valid paper route and you still won’t get your compound synthesized. Where is that salient little point in this exercise? If I were one of those 12 contestants, I wouldn’t want Beevis and Butthead at some CRO that we all know is a cheap substitute for what used to be called a research lab trying to work out my route. That’s for sure.

    1. Kazoo Chemist says:

      That was my initial thought also. Will all routes be evaluated by the same set of “hands”?
      If not, then forget about the statistical evaluation of the proposed routes, the largest variable will be at the bench.

      What level of effort will they make to evaluate and optimize the given route. There have been many posts on this very blog about using brute force multivariate methods development to explore seemingly “routine” reactions such as a Buckwald amine coupling. These show very convincingly what practicing chemists know all too well: minor changes in conditions (solvent, temperature, catalyst, added ligands, etc., etc., etc.) can have a huge impact on the outcome of a reaction. Will the “contract synthesis labs” be spending any effort to optimize any of the steps, or will they run them as a one-off using the exact conditions presented by the submitters? If “Beevis and Butthead” (great choice for their moniker!) pull some boner move and screw up a reaction will that be the end of the submitter’s chance for glory? I could be wrong (I often am) but from what I have read this is a very poorly designed experiment that should yield a correspondingly poor result. A well intended idea, but there are far too many issues to make this a truly valid comparison of twelve proposed synthetic routes. If one were to be very cynical one would suggest that this sort of smells like a beauty pageant where the local favorite has already been fitted for the tiara. I look forward to seeing the list of “chosen ” competitors, and the final list of the “winners”. I would likewise enjoy seeing the list of entrants who were not chosen for the final twelve.

    2. Ru barf says:

      Well, did you ever consider that if your route can only be run by such a ~TaLeNtED~ *thought leader* then maybe it ain’t such a good route after all!!

    3. Why the high horse? says:

      ‘I wouldn’t want Beevis and Butthead at some CRO that we all know is a cheap substitute for what used to be called a research lab trying to work out my route’

      I find this highly insulting. I’ve worked for some time at a large drug discovery CRO and many of their chemists are highly talented and by comparison, often out perform big pharma chemists who get cosy behind desk jobs.

      1. anon2 says:

        Not on a high horse at all and agree that there are some very talented chemists at CROs and certainly some of the best are employed there. Let me qualify what I said so you understand: I would not want Beevis and Butthead at any company on the planet trying to work out my route. The point is that if you’re using a desk jockey or a computer to generate your route, then you don’t need internal chemists and will therefore likely outsource that work, hence my reference to using B&B at a CRO. It is my opinion that we should not assume that all CROs or the chemists within are created equal by any measure. The same applies to big Pharma, small Pharma, etc. It was not a dig against CROs in general. It was to point out that not all chemists are created equal with the same paper route. Feel better? BTW, I worked as a chemist at a medium size CRO. I got nothing against CROs.

  5. Wavefunction says:

    I hope they do the statistics on the results right. In that Kaggle challenge from a few years ago, the difference between the winning and second (and perhaps third) entries was statistically indistinguishable, yet the first entry was declared as the winner.

  6. Magrinho says:

    I just had a dream that it was 1995 and people were talking about how to make molecules.

    It was a very pleasant dream…

  7. anonymous chemist says:

    I sincerely hope no morons take part in this exploitative “crowdsourcing” challenge, which is nothing more than a sneaky way to save money. Merck has plenty of money to hire chemists and actually pay them for their work and supplies. A single $10k payment for solving a synthetic route is paltry considering the resulting IP will belong to the company. An above anonymous poster eloquently put it, there are many hidden costs to participating in this moronic challenge, such as literature and database access, not to mention the time and effort involved.

    If this thing takes off, you will absolutely see more and more of these crowdsourcing challenges. The next step for these companies is turn the experimental side into the “gig economy” by outsourcing the synthesis of chemical intermediates to the crowd.

    “Make this 5g of this heterocycle and get a chance at winning $500!” – There Merck, I came up with your next big idea.

    1. Derek Lowe says:

      You could take a look at the site and notice that all the IP stays with the submitter, and none with Merck.

      1. A Nonny Mouse says:

        Probably looking for a cheaper route to praziquantel as the price has jumped from about $80/kg to $140 as the Chinese people using the Reissert process have shut down due to pollution with cyanide. Everyone has now moved to the more expensive Shin-Poon route.

        As Merck gives this away for free, their costs have substantially increased.

        If they asked me, I could give them 2 (reduced to practice as well)!

      2. anonymous chemist says:

        Fair enough. I missed that one and you’re right to point it out.

        But that doesn’t invalidate my other points regarding these types of crowdsourcing challenges. Merck benefits disproportionately from the efforts of participants. They do not have to pay for chemists for their time, their database/journal access, nor their supplies. I wouldn’t be surprised if the use of academic journals and databases for commercial efforts like this violate the TOS.

        If this crowdsourcing trial run proves successful, there’s no telling if future challenges wouldn’t have more restrictive rules regarding intellectual property. And even if the IP were to remain with the author in future cases, its going to be individuals vs large corporations, who always have deeper pockets to fight legal battles. Patent evasion has existed since there have been patents to evade, and pharma companies have extensive experience with this.

        Sorry Derek, but I can’t see how these crowdsourcing efforts, aside from providing an entertaining intellectual challenge, benefit the regular chemist.

  8. anon the II says:

    This is the stupidest thing I’ve seen in a while. First the web site hits you with a massive survey in the first millisecond of the page loading. And I’m stupid enough to fill it in just so I can have the opportunity to tell them how stupid it is to do things like that. And then why would I want to go to the trouble of filling out the form. What chance in hell do I have of getting selected? Why don’t they just show us the god-damn molecule and let us have a go at it. Why don’t they just call up Baran and Nicolau and a few more big names rather than trying to make a whole bunch of people feel insignificant because we didn’t get chosen. Jeez!

    1. Anon says:

      “Why don’t they just call up Baran”

      He probably charges more than 10k.

  9. David says:

    This looks like a great competition that works for the sponsor and the entrants. Entries remain confidential and IP remains with the entrants, so you can’t be too cynical about Merck KGaA’s motives. Great outreach and a good opportunity for students.

  10. Anon says:

    10,000 Euros / 96 hours / 12 entrants = 8.68 Euros per hour per entrant (gross average, excluding costs).

    Any jobs going at McDonalds?

  11. Anon says:

    When Innocentive was launched internally at Merck, nearly all of the “winners” were direct reports of the judges, or judges themselves. The next year, hardly anyone bothered to submit.

  12. Andy says:

    “Merck KGaA, Darmstadt, Germany will provide the structure of a small molecule to all selected participants at the same time.”

    Does that mean it’s the *same* small molecule to each participant?

  13. Mister B. says:

    Dear all readers,

    We’ve gathered a team of 6 Ph.D students and post-doc and we will apply for this challenge. No matter what you think about the hidden reasons from Merck, it is still a good chemistry exercise and whatever happened we could use our brain and share ideas.

    Thank you Derek for sharing this ! Enough bad thoughts and cynicism !

    1. anon the II says:

      You’re right. I apologize for being cranky. Maybe it was that stuff going on in Helsinki that had me in a bad mood. Still, some of us old one-offs, put out to pasture before we were ready, would like a crack at going head-to-head with the A-teams.

    2. Anon says:

      Let me know if they also have time to do my laundry.

  14. Brendan Monks says:

    Hi All,

    I am part of the team that is setting up this challenge at Merck KGaA, Darmstadt, Germany. Thank you Derek for posting the challenge and all the exciting conversations it is fostering. I just wanted to jump in and answer a few questions I saw in the reply thread.

    -The main motivation for launching the challenge is the celebration of our companies 350th anniversary
    -All selected participants will get the same molecule to synthesize
    -Selected participants have 96 hours to propose a synthesis, they certainly don’t need to take the entire time and can submit their proposal when they think they are complete
    -No one from Merck KGaA, Darmstadt, Germany or an affiliate is allowed to compete in the challenge
    -As mentioned reply thread, entries remain confidential and IP remains with the entrants
    -The challenge is not associated with Innocentive in any way

    Thank you again for your enthusiasm. If you have any questions please do not hesitate to contact me directly at: open.innovation@emdgroup.com

    1. Anonymous says:

      There are too many unknowns, but I’ll just blurt out these thoughts.

      1. Make it TRULY open, either before the fact or, even, after the fact. Is the small molecule a proprietary structure or just a test of Chematica vs humans? Pick something challenging but non-proprietary so it can be truly open.

      I wouldn’t be surprised if Pipeline contributors could offer up some challenging suggestions.

      2. After the “closed” (12 only) judging is over, open it up to complete peer competition (more ideas) and peer review (e.g., in a blog such as Pipeline). In the current format, confidentiality is guaranteed. Assuming that some of the 12 will disclose their answers, let the entire community offer their opinions. Or offer another challenge on a completely non-confidential format.

      3. Pipeline contributors have often complained about the lack of originality in “original” papers. I have refereed (rejected) papers that were tantamount to plagiarism of previously published work, w/o proper attribution. (No one famous. Editors stopped sending me mss after a few of those.) There are famous cases of proposals that were plagiarized (Paquette / Holton / Taxol) as well as many w/o such fame or notoriety.

      In Pipeline, “Precedential Citations” 9 March, 2018 at 3:23 pm commented, q.v., that Gryzbowski did not give adequate credit to Hendrickson’s Syngen work on linking synthetic routes to literature precedents. Merck judges might not have known about that w/o opening up review to a wider audience.

      4. If your CRO cannot get Entry # to work, I would not be surprised if talented voices on Pipeline could point out quick fixes or clever solutions. (I was with a biotech that outsourced a project to a CRO. They failed to deliver on time and the project went south. I was sent to look over their work. The notebooks actually showed how they had failed to follow detailed instructions in the provided experimentals. [“Intermediate X is unstable and must be immediately combined with reagent Y. Only then can it sit overnight.” Repeated failures, documented in the notebooks, because they let X sit overnight! Unbelievable. Many chemists here on Pipeline would know that X is unstable even without the boldfaced written instructions.])

      5. I’ve got some other “innovative” ideas. Perhaps change the name from “open.innovation” to “partially.open.innovation” or “selectively.open.innovation”.

  15. Anonymous2 says:

    Just wondering as to those blogs that revolved around (retro)synthesis of Natural products…such as “totally synthetic” , “totally retrosynthetic” and “Just Like Cooking” and hope they comeout of dormancy to blog/participate in discussions/contests.

  16. Anonym says:

    Just wondering as to those blogs that revolved around (retro)synthesis of Natural products…such as “totally synthetic” , “totally retrosynthetic” and “Just Like Cooking” and hope they comeout of dormancy to blog/participate in discussions/contests.

Comments are closed.