B.R.S.M. The road to Tet. Lett. Is paved with good intentions


And Now For Something Completely Different 1: Progress In Strychnine Synthesis

Sorry things have been so quiet around here the last couple of weeks - I've been a bit uninspired, the start of term has been a bit hectic, and my boss has finally noticed that I haven't really made any progress in the last six months. Hopefully more things soon as I wrestle my life back under control.

A few weeks ago, as the result of a conversation in the lab, I thought it might be interesting to pick a popular molecule (i.e. something of which there have been numerous syntheses), and see how much better we've gotten at making it over time. After a bit of thought I settled on strychnine, and in order to not waste an excessive amount of my life, only took syntheses which produced a single enantiomer of the natural product. So, first let us look at how the number of steps (in the longest linear sequence) has varied over the years.

Right, so that's kind of as it should be - not the sharp downward trend one might have expected, but at least it's decreasing over time. Why, by 2110 we'll be making it in no steps at all! And it'll make a good undergraduate experiment in eighty years or so.[1] Excited and thrilled by my predictive powers (and being a fairly cynical person) I thought it'd also be interesting to plot the average yield per step against year.[2]

Huh, so it turns out that we're not really getting any better yields now than R. B. Woodward was half a century ago. And no perfect synthesis for a few thousand years, if present trends continue. Personally, I'd expected a much greater upward gradient for this graph, as we, err, become better chemists. Or something.


1. Probably. I failed AS maths.

2. I found synarchive.com to be a massive help in this exercise. Where yields were only available over 2 steps (or more; fairly common in the literature) then I've just counted the 2 steps as 1 with the yield given. I realise this is a very stupid thing to do, but this is not a serious post, and I can't think of a better way to fill in the gaps where the yield for each step separately simply isn't available. Also, average yields were worked out using my phone, while watching old episodes of Red Dwarf. So I wouldn't publish this data without checking it.


Comments (6) Trackbacks (0)
  1. Interesting and slightly shocking post there. Who’d have thought that the yields aren’t getting any better?! And it is not as if chemists lost interest in the natural product there were quite a few attempts in the last decade. How weird.

    P.S. Why strychnine?

  2. You know, I’m not really all that surprised. I think its more of a human thing thats being reflected. People won’t publish unless the overall yield per step is respectable (meaning the overall yield of the target is good) but they also won’t waste too much time getting every single yield >90%. Very cool though that syntheses are getting short (hopefully meaning better)

    @ AzaPrins Because strychnine is hardcore 😛

    • I definitely think people do much more optimisation if the yield is below 60% to try and get back in the 80-90% range, but diminishing returns often set in for >90% where you need more and more time to get those last few percent. You could look at the data and say that everyone’s prepared to put in a similar amount of effort up to a certain point, explaining why they’re all similar. This does hide some fairly large variation, though – Woodward’s synthesis has 4 awful steps in the 12-30% range, which you rarely see nowadays (maybe more so in the first synthesis of a hot target) where as Shibasaki reports three consecutive steps with 100% in the middle of his route (which really isn’t possible after 3 workups + columns) and yet the averages are extremely similar.

      I expected yields would trend upwards as people started to be under more pressure to publish in ‘higher quality’ journals and maybe exaggerate a bit. We’ve all heard stories of postdocs being told ‘you take care of the reactions, I’ll take care of the yields’ by their professors before, and certainly if you read anything Tomas Hudlicky has written (namely The Way of Synthesis, or that piece he wrote in Synthesis last year) you’d think the problem of yield inflation was fairly widespread.

      @Aza – Yeah, I was surprised too. I expected this to be a rant on the increase of average yields, which it looked like for a while, but then when I rechecked everything the line became almost flat. I chose strychnine because I wanted something with a lot of syntheses over a long period of time (although there’s a huge gap between Woodward and the rest). Other contenders included kainic acid, salinosporamide and a few other (indole) alkaloids. But why not pick a target everyone knows? Also, Synarchive contains all of the strychnine syntheses, making the number crunching a bit easier. And ckellz is right.

  3. Where N steps are grouped together just take the Nth root of the overall yield (i.e., if the yield for 2 steps is 81% (.81) assign each step a yield of 90% (.9; .9 x .9 = .81), 50% over 4 steps is 84% each (.84 x .84 x .84 x .84 = .5)).

  4. While the Woodward synthesis does make single enantiomer of strychnine, it technically shouldn’t be classified as a single enantiomer synthesis because it is a relay synthesis (actually a double relay)

Leave a comment

No trackbacks yet.