from Hacker News

We Made One Gram Of Remdesivir

by tomstokes on 4/5/20, 2:49 AM with 229 comments

• by kccqzy on 4/5/20, 4:39 AM

  Remdesivir is an analog of adenosine, one of the four building blocks of RNA. Just look at the main structure and you'll agree they look similar. It turns out the mechanism of action of this drug is that it's supposed to be confused with adenosine, so that the viral RNA replication process uses remdesivir instead of adenosine, which later breaks the RNA†.

  Our body, or really, all biological processes can synthesize incredibly complicated molecules that can take human chemists a huge amount of effort to synthesize. It really is amazing how awesome our body is.

  †: My description here is a dumbed down description. For a more precise description see section 2 of Arguments in favour of remdesivir for treating SARS-CoV-2 infections, Wen-Chien Ko et al, https://www.sciencedirect.com/science/article/pii/S092485792...

• by eyegor on 4/5/20, 4:31 AM

  This article really reinforces my choice to drop ochem in college. It's funny how deciphering assembly code seems downright mundane compared to this jargon. Like, I get how given the process you could recreate things, or you could devise possible reactions based on electron shells and bonding tendencies, but how do researchers even figure out how to make these reactions happen? Especially when they require -100C, or +5atm of pressure.

• by aerovistae on 4/5/20, 3:33 AM

  Really interesting getting a window into the world of professional chemistry. So impactful on all of us but so much more opaque. We see CS jokes/memes all over the place, even if only those who program get them, but we never see chem jokes, or only sparingly and rarely beyond a high school level.

• by Vaslo on 4/5/20, 5:09 AM

  In my former life I was a graduate organic chemist, but I “quit” with a Masters. I loved Synthetic Chemistry as an undergrad and thought I wanted a PhD in. There were three lousy things (for me) about the career that made me bail.

  1). So many syntheses have horrible yields just like this one. You’d start with grams of material to end up with micrograms. I loved solving these problems as an undergrad in books, but reality was far different. You don’t think much about side products until you start doing novel chemistry.

  2). So much trial and error. There were happy go lucky chemists that fell into projects that were smooth as butter, while brilliant chemists would toil 12 hour days to try and something to write up as a thesis. I was neither brilliant nor lucky and took 4 different projects over two years before finally landing on something marginally MS worthy. They need a journal of failed chemistry because only the working stuff gets published. So many failures could be logged so I didn’t waste my time doing non-working or poor yielding reactions.

  3). Suspicious results in journals. I would read about a reaction and someone would put a 75% yield as their result and I could barely get 20. I always thought I was just bad, but a really smart chemist challenged me one day and tried to do it himself and couldn’t do much better. He tried it 30 different ways over the year as he did other stuff. He never could get a good yield. We talked to our advisor and we wanted to challenge the result, but the advisor didn’t want to start trouble. It was past the time I decided to leave with a masters, but made me feel a little better about my lousy abilities. No one could ever possibly doublecheck every result from every publication anyways.

  All this said, there are some brilliant and patient scientists out there that drive the field forward. Just a few rough around the edge items I’d love to see change.

• by est31 on 4/5/20, 8:02 AM

  > Here are the real numbers for the reactions in Scheme 1: 0.25 x 0.58 x 0.74 x 0.21 x 0.23 = .005 (0.5%)

  This reminds me of the problems to scale up EUV lithography which are bottlenecked on producing strong enough EUV light. They put in 20 kW of power to get out 200 W at the target wavelength of 13.5 nm, so light generation itself only has 1% efficiency, and then you need to reflect it at mirrors etc. to focus it (lenses don't work at those wavelength) and that makes only 2% of the light actually reach the waver [2].

  [1]: https://www.laserfocusworld.com/blogs/article/16569161/the-s...

  [2]: https://en.wikipedia.org/wiki/Extreme_ultraviolet_lithograph...

• by ebg13 on 4/5/20, 3:34 AM

  The writing is so good, I had to double check that it wasn't written by Derek Lowe.

• by tikej on 4/5/20, 12:37 PM

  I want to make remark about organic synthesis from the perspective of person who is engineer of chemistry by training as did his fair share of organic synthesis (and eventually moved to computational atomic and molecular physics).

  It is one of the oldest and very established fields. Unfortunately practices aren’t great. The preparation formulas are often vogue, imprecise and difficult to reproduce. This comes from the fact that often the sizes and types of glassware are not specified, some informations are omitted (how quickly something is changed not only to what value e.g. heat up to 100 degrees but it does not say over what time) etc. Chemists usually (except some theoretical/computational specialisations) don’t have any training in algorithms or programming.

  There are novel developments such as https://www.gla.ac.uk/news/archiveofnews/2018/november/headl... and references therein. I’m optimistic about them but I expect strong opposition from older faculty. They see synthesis as more than art and think that one has to have “good hand” in order to be a good organic chemist.

  I think some generational shift will be necessary in order to change this discipline to more reproducible, strict and reliable. It will come but not that soon :)

• by franciscop on 4/5/20, 3:54 AM

  One too many zeroes here:

      0.25 x 0.58 x 0.74 x 0.21 x 0.23  = .005 (0.05%)
  

  The 0.005 seems to be correct, so that should be 0.5%. The rest of the article also uses "0.5%" correctly.

• by mettamage on 4/5/20, 4:50 AM

  This was a lot of fun to read and quite accessible. I remember that I was horrible at chemistry in high school, but the right things were ELI5'ed.

  I like HN in general, but this particular article gave me the same feeling that I had when I first discovered HN a couple of years ago.

• by roenxi on 4/5/20, 4:26 AM

  Although very interesting, this article isn't a great introduction because a critical part of the picture is missing - the availability of the inputs.

  Sticking to simple stuff; mines produce iron at a rate measured in thousand-tonnes per hour with yields of potentially sub-30% compared to volume of earth moved. Ammonia and many acids are presumably measured in tonnes or kilograms produced per day. Low yields make the process-oriented sad, but what matters is absolute ability to produce; not yield.

  All that doesn't take anything away from this article; it just makes it hard to interpret what 'royal pain to synthesize' means in practice. The process isn't basic chemistry; but that isn't really saying much.

• by endorphone on 4/5/20, 11:12 AM

  Chemistry is absolutely fascinating. One of my long-time goals has been to make a learning chemistry game of some sort. Players building bonds and molecules, solving problems by creating reactions, etc. Orbits and ions and quantum states.

  Something truly educational, and of course one of the people I wanted to educate was myself by forcing myself to learn much more than my high-school level of chemistry.

  I've gone down this road a few times, each time giving up on the absolutely scale of even the most rudimentary understanding.

• by abhisuri97 on 4/5/20, 4:48 AM

  This brings me back to orgo lab. A six step synthesis was torture when you were graded by your final yield exactly bc of the compounding nature of small mistakes in each step. Even if you get fairly good yields of 80 to 90% per step, you’d end up with a 53% at best and hope to god that everyone else did equally badly so the curve would save you.

• by unnouinceput on 4/5/20, 9:20 AM

  Quote: "Fortunately, there are creatures called "process chemists" – guys who laugh at the discovery synthesis, insult as many of us as possible, and then make it better so that the yields are higher and some of the most dangerous chemical reagents and solvents – ones you really don't want to use on a large scale – are replaced by others. They are generally a bunch of mutants, but they do a great job in "fixing" the synthesis used by the discovery chemists."

  This really cracked me up. In my IT world the equivalent for "mutant" would be refactoring, right? I did "mutated" this way a few times in the past to much of the horrors of my boss(es)/manager(s) when they learned the next day.

• by bionhoward on 4/5/20, 11:45 AM

  ML is useful to design the reaction pathways to make these chemicals. If we find safer/better/cheaper ways to make medicines like Remdesivir with deep learning, and publish them, then chemists may have an easier time of it.

  If Remdesivir data looks good this month, there will be a rush to produce it, and if there’s only one published way to do that, then the ingredients for that one approach will potentially be hard to find. Thus we can benefit from different approaches which start from different raw materials.

  Lots of cool Arxiv papers on this and Graph Neural Nets, Soft actor-critic, or Transformers can be interesting approaches. The transport theory seems like a good way to make a value function. How much time and money does it take to produce a given chemical by a given set of reactions? That’s a gajillion dollar question.

  I spent way too much time last year looking at permutation-invariant distance metrics similar to Fused Gromov Wasserstein to invent an Atom Mover Distance, please let me know if you figure that out! DeepChem library is a solid framework, as are Tensorflow and Pytorch...

  If anyone’s looking for a way to contribute to the COVID-19 response, open source data/algorithms to design synthesis pathways can be a strong approach. Everyone loves to use Deep Learning to design drugs, but it is valuable to design ways to make drugs, too!

• by gorgoiler on 4/5/20, 4:19 AM

  Very funny writing style while being incredibly informative. Read from top to bottom with glee. I had a feeling organic synthesis was difficult but not this difficult.

  Missed a trick not titling it:

  “(1OO)OMG we made one gram...” :)

• by failuser on 4/5/20, 8:29 AM

  This makes the biological chemistry so awe-inspiring. Everything is done at about the same body temperature and at the same pressure, just with right catalysts.

• by manav on 4/5/20, 3:34 AM

  It’s been almost two decades since I took a chemistry class, looks more foreign to me than code to a non-programmer.

• by saadalem on 4/5/20, 1:49 PM

  I was able to create several small molecule candidates which achieved docking scores of almost -18 but they still need to be assessed for their synthetic feasibility. (Docking scores of leading existing drugs (HIV inhibitors) are around -10 to -11 (the more negative the score the better), Remdesivir drug is around -13 which already entered clinical testing.

  I know that binding affinity has been shown not to be the best indicator of efficacy always, but I want to know if it's feasible, if someone can help

• by jluxenberg on 4/5/20, 9:35 AM

  Press release from Gilead published yesterday says they aim to have 500,000 treatment courses by October (they reduced a 12 month production process to 6 months): https://www.gilead.com/stories/articles/an-update-on-covid-1...

• by VSerge on 4/7/20, 8:03 AM

  Oh the pain of biochemistry! Thanks to the author for making this available to everyone, and reminding these of us who ran screaming from the labs why our decision is not among the ones we might regret. Joking aside, I hope this molecule can be synthesized efficiently, and that it can be preempted from Gilead somehow, as this company has a history of price gouging of the worst kind. They are down there with the oxy makers, the worst of the worst, causing deaths through their pricing policies, and resulting lack of affordable treatment. The ideals of medicine mean nothing to these, and public health should take precedence over the interests of such despicable companies.

• by ngngngng on 4/5/20, 5:12 AM

  I've never learned a bit of Chemistry. I assume there are formulas to determine how substrates and reagents will react with each other? And the process chemists understand these formulas better than most?

  Apologies for the assumptions in these question, but are there many reactions in organic chemistry that are completely unknown?

  This actually seems pretty fun. I'd love to have a reason to study it and a means to do something with my studies.

• by garmaine on 4/5/20, 7:21 PM

  > remdesivir, possibly the most promising candidate

  Is there any evidence of this? It was a highly anticipated drug, but the first studies it was in showed only slight improvement over expected outcomes, far less than was seen with the chloroquine/zinc/antibiotic combo treatment.

  I mean I think it's still interesting as a possible drug to add to the cocktail for maximum effectiveness. But let's not oversell it.

• by joyj2nd on 4/5/20, 11:43 AM

  "liquid chromatography, is frequently used. On a small scale this is fine, but when kilograms of material are required the process becomes unwieldy."

  I did not understand this. liquid chromatography scales up quite well. There are other methods like Electorphoresis, salt precipitations etc., that don't.

• by aazaa on 4/5/20, 5:12 PM

  > n-BuLi is short for n-butyllithium, which bursts into flame if exposed to oxygen or water.

  It looks like the author is overselling some of the dangers here.

  While you really don't want to dump n-BuLi into water, you have no reason to either.

  The problem child of the class to which n-BuLi belongs is t-BuLi. That will spontaneously ignite in air, whereas n-BuLi will not. There was a very high-profile case I believe at UCLA a few years back in which a student using t-BuLi in the lab caused a fire with it and ended up dying.

  https://cen.acs.org/articles/87/i31/Learning-UCLA.html

  Also, I find this article confusing in the way it's written. Take the title, for example. It gives the impression that the author is describing his own efforts to make remdesivir ("we").

  What he's really describing is some preps he found in the literature. And with a little too much hyperbole for my taste.

• by Scoundreller on 4/5/20, 1:32 PM

  "(2) The one-gram dose is a guess based on the doses of other antiviral medicines."

  You're in luck, the first dose is 200mg, and then 100 mg daily after that.

  Unfortunately, it's also IV, so you have a number of extra steps after synthesis to ensure sterility.

• by aledalgrande on 4/5/20, 3:34 AM

  LOL at the way process chemists were introduced

• by jijji on 4/5/20, 6:49 PM

  favipiravir is alot cheaper to synthesize and is used as a drop in replacement for remdesivir all over asia...

  Good things to have on hand during this covid-19 pandemic (because you can't rely on hospitals to give it to you) are:

  1) hydroxychloroquine sulfate taken orally 400 mg per day for a week ($180/kg on alibaba)

  2) azithromycin taken orally 500 mg per day for a week ($150/kg on alibaba)

  3) Camostat mesilate taking orally 200 mg three times a day for a week [1] ($50/g on alibaba)

  4) favipiravir one dose of 1600mg two times on the first day, and then 600mg twice per day after that for a week [2]. ($40/g on alibaba)

  5) covid-19 rapid test kits that use blood antibody tests and produce results between 3 to 10 minutes and cost about $1.50 per kit [3] .... although it looks like in the last week or so Alibaba has been blocking searches for these kits for some reason... although the search below does work but you have to look at the suppliers to find the ones that are actually selling it.

  all this stuff can be bought on Alibaba and delivered in a week

  [1] https://clinicaltrials.gov/ct2/show/NCT04321096

  [2] https://www.medicalnewstoday.com/articles/anti-flu-drug-effe...

  [3] https://m.alibaba.com/products/covid-19_rapid_test_kit.html

• by fg6hr on 4/5/20, 4:05 AM

  That's some really complex mess. How much do these chemists get paid?

• by Josh_Bloom on 4/5/20, 5:14 AM

  @ebg13 Thank you! That may be the biggest compliment I’ve ever gotten. Derek is the god of carbon as far as I’m concerned. Josh

• by Glosster on 4/5/20, 1:34 PM

  Does anyone know what's the probability that Remdesivir is indeed a solution against Coronavirus?

• by franciskim on 4/5/20, 9:11 AM

  Strange that there's not one mention of Avigan / Favipiravir.

• by quickthrower2 on 4/5/20, 7:11 AM

  Funny just saw this comment on reddit. Related https://www.reddit.com/r/Political_Revolution/comments/fuuuy...