Skip to main content

Drug Assays

Different Drug Companies Make Rather Different Compounds

Now here’s a paper, packed to the edges with data, on what kinds of drug candidate compounds different companies produce. The authors assembled their list via the best method available to outsiders: they looked at what compounds are exemplified in patent filings
What they find is that over the 2000-2010 period that not much change has taken place, on average, in the properties of the molecules that are showing up. Note that we’re assuming, for purposes of discussion, that these properties – things like molecular weight, logP, polar surface area, amount of aromaticity – are relevant. I’d have to say that they are. They’re not the end of the discussion, because there are plenty of drugs that violate one or more of these criteria. But there are even more that don’t, and given the finite amount of time and money we have to work with, you’re probably better off approaching a new target with five hundred thousand compounds that are well within the drug-like properties boxes rather than five hundred thousand that aren’t. And at the other end of things, you’re probably better off with ten clinical candidates that mostly fit versus ten that mostly don’t.
But even if overall properties don’t seem to be changing much, that doesn’t mean that there aren’t differences between companies. That’s actually the main thrust of the paper: the authors compare Abbott, Amgen, AstraZeneca, Bayer-Schering, Boehringer, Bristol-Myers Squibb, GlaxoSmithKline, J&J, Lilly, Merck, Novartis, Pfizer, Roche, Sanofi, Schering-Plough, Takeda, Wyeth, and Vertex. Of course, these organizations filed different numbers of patents, on different targets, with different numbers of compounds. For the record, Merck and GSK filed the most patents during those ten years (over 1500), while Amgen and Takeda filed the fewest (under 300). Merck and BMS had the largest number of unique compounds (over 70,000), and Takeda and Bayer-Schering had the fewest (in the low 20,000s). I should note that AstraZeneca just missed the top two in both patents and compounds.
radar plot
If you just look at the raw numbers, ignoring targeting and therapeutic areas, Wyeth, Bayer-Schering, and Novartis come out looking the worst for properties, while Vertex and Pfizer look the best. But what’s interesting is that even after you correct for targets and the like, that organizations still differ quite a bit in the sorts of compounds that they turn out. Takeda, Lilly, and Wyeth, for example, were at the top of the cLogP rankings (numberically, “top” meaning the greasiest). Meanwhile, Vertex, Pfizer, and AstraZeneca were at the other end of the scale in cLogP. In molecular weight, Novartis, Boehringer, and Schering-Plough were at the high end (up around 475), while Vertex was at the low end (around 425). I’m showing a radar-style plot from the paper where they cover several different target-unbiased properties (which have been normalized for scale), and you can see that different companies do cover very different sorts of space. (The numbers next to the company names are the total number of shared targets found and the total number of shared-target observations used – see the paper if you need more details on how they compiled the numbers).
Now, it’s fair to ask how relevant the whole sweep of patented compounds might be, since only a few ever make it deep into the clinic. And some companies just have different IP approaches, patenting more broadly or narrowly. But there’s an interesting comparison near the end of the paper, where the authors take a look at the set of patents that cover only single compounds. Now, those are things that someone has truly found interesting and worth extra layers of IP protection, and they average to significantly lower molecular weights, cLogP values, and number of rotatable bonds than the general run of patented compounds. Which just gets back to the points I was making in the first paragraph – other things being equal, that’s where you’d want to spend more of your time and money.
What’s odd is that the trends over the last ten years haven’t been more pronounced. As the paper puts it:
blockquote>Over the past decade, the mean overall physico-chemical space used by many pharmaceutical companies has not changed substantially, and the overall output remains worryingly at the periphery of historical oral drug chemical space. This is despite the fact that potential candidate drugs, identified in patents protecting single compounds, seem to reflect physiological and developmental pressures, as they have improved drug-like properties relative to the full industry patent portfolio. Given these facts, and the established influence of molecular properties on ADMET risks and pipeline progression, it remains surprising that many organizations are not adjusting their strategies.

The big question that this paper leaves unanswered, because there’s no way for them to answer it, is how these inter-organizational differences get going and how they continue. I’ll add my speculations in another post – but speculations they will be.

30 comments on “Different Drug Companies Make Rather Different Compounds”

  1. Lacerta Bio says:

    Wow. It would be interested if the data could somehow be corrected for therapeutic area or target focus. Is Vertex “better” because of their focus on anti-virals?

  2. Anonymous says:

    according the paper the same inter-compony trends persist even after adjusting for target class.

  3. CMCguy says:

    I agree that once get to clinical candidates it is most critical to be more “drug-like” however not so sure that would go as far in innovative discovery with “you’re probably better off approaching a new target with five hundred thousand compounds that are well within the drug-like properties boxes rather than five hundred thousand that aren’t.” A “good mix” might be more fruitful. Yes that would be preferred if possible to have inputs that are closer to drugs but what is most important is to gets relevant “hits” to start a project which can turn to med chem leads which then can design to candidates. So even though would populate screening library with majority drug-like compounds would advocate not restricting all to such criteria in attempts introduce molecules that could if lucky be initial hits even though further from the goal. Of course its important to not stock with overly active functionality that waste time with false positives. Perhaps the only perfect libraries are ones that appear in publications after the fact.

  4. cynical1 says:

    Well I hope that database they bought to do this analysis distinguishes between the final analogs in the patent as opposed to all the intermediates that are reported in the patent to make them. I’m pretty sure that my pentultimate trityl-protected amine is going to have a high cLog P. So what? What if I had to make 50 separate trityl or FMOC protected amines?
    If they didn’t separate intermediates from analogs (and it wasn’t clear to me from their website that they do), then you’re probably looking at a futile exercise that probably more accurately reflects which companies are willing to undertake multistep synthesis to tease out the SAR. My two cents….

  5. I think Derek makes a great point. I know that IP strategies can vary a lot between companies. When I first starting working in the pharma industry, the company I was working for patented almost anything of value. Once we were acquired by Pfizer, it was hard to get anything patented. So a lot of programs that stopped in either the pre-clinical stage or early clinical stage never ended up getting patented.
    That would obviously give a very biased view of how Pfizer’s med chem design efforts fit into this graph.

  6. completelyfedupwithpharma says:

    Although my real work observations in a big pharma medicinal chemistry department tend to be in line with the paper’s analysis (I am working in one the pharmas which scores worst) I also have to admit that I had similar thoughts as cynical1. Patent DBs I have worked with were usually quite contaminated with intermediates and whatever chemicals. Should however be about the same for any company’s patent and thus the effect probably averages out….

  7. SwedenCalling says:

    Please explain to me why *all* medchemist trust calculated clogP values? Most other models are usually consider unreliable. But clogPs are holy. Odd to me. The error in the logP predictions are not seldom more than 2 log units…in either direction. This together with all other uncertainties mentioned in comments above makes such a comparison study useless, at best.

  8. barry says:

    it would mean more to break it out by target, rather than therapeutic area. A group working on neuraminidase inhibitors is going to submit low-molecular weight compounds with very low logPs. A group working on HCV protease inhibitors will mostly submit compounds of much higher molecular weight and logP. Yet these authors will call them both “anti-viral”.

  9. DrSnowboard says:

    Are Leeson et al ever going to stop banging on about this one idea and actually discover something? As in a drug?

  10. Anonymous says:

    @4 I think they do, because they will use the GVKBIO patent database as before.
    Interesting would be an additional graph with launched drugs for these companies in about the same time line – the ones which came from internal discovery (no merger, buy-in etc).
    @7 calculated logP values can be expected to be 0.5 log units off – but the comparison depends also on the experimental method used to determine logP. In this respect I never understood why shake flask is so prefered.

  11. Anon says:

    Seems like Nature bla bla accepts everything nowadays. The only good thing coming out of Leesons work is his AZ colleagues ironic graph in a recent DDT paper.

  12. Ed says:

    #9 Ouch! No doubt they will claim partial ownership of any AZ internal pipeline success stories due to their strategic input! But I totally agree – the amount of effort that goes into this kind of work over the years is huge and it is hard to understand the justification for its continuation.
    If I remember right, AZ has had one of the least productive pipelines over the past ten years: a string of high profile Phase 3 failures. Obviously the efficacy part of the analysis has yet to be optimized. Should keep them busy for a another five years

  13. Hopefully no evidence of Categorical Sin

  14. TX Raven says:

    @ SwedenCalling: Not “all” chemists do. I agree with you that cLogP is a very poor predictor of LogP, in particular when you look at novel heterocyclic compounds.
    By the way… have you seen the “putative” structure of the mGlu5 NAM that Roche came up with? Horrible in terms of pchem properties, but it is kicking ass BIG TIME in the clinic in terms of daily dose…
    Keep pouring the Lipinski kool-aid…

  15. barry says:

    Lipinski has always made clear that these are correlations, not laws, and that they apply only to passive transport. Whole classes of compounds (penicillins…) have had glorious careers outside the Lipinski box.

  16. NRDDesist! says:

    Sorry but does this actually contribute anything of use?
    I despair of the trend for articles like this whose sole purpose seems to be to provide a title capable of luring in the $32 it costs to read them.
    I suspect most practising Medicinal Chemists will treat this with the scepticism it deserves. Unfortunately, their managers may seize upon it as proof their organisations are underperforming and initiate yet another dim but damaging initiative to remedy “their” problem.
    Messr’s Leeson and St-Gallay, this is not helping.

  17. UK Chemist says:

    #3 Exactly right , if you’re going to bet on the horse and keeping betting on the long shots some of you will win big ,but most will go bust. Drug space is just an area where there are likley to be more oral drugs and if Mgt actually understood that we might be more successful. It just looks like any other are of conditional probability

  18. Anonymous says:

    Did anyone else notice the big difference between the cLOGp values of Pfizer and Wyeth compounds. Perhaps that is why Pfizer pfired all the Wyeth chemists? It could be a difference in patenting strategy (patent late vs. patent early) although I think Pfizer changed their approach half way through the time period under study.

  19. And D says:

    #16, #9,
    Surely investment into understanding why attrition is high has got to be worthwhile. How to practically use the information to devise molecules with a better chance of success is where things get controversial.

  20. MedChem says:

    #16, 9, 20
    One of the first things any medicinal chemist learns is “high cLogP is bad!”. Very rudimentary and simple really. It’s people who are incapable of creating anything of value that are making a career out of banging on this stuff over and over and over and over again.

  21. MedChem says:

    #16, 9, 20
    One of the first things any medicinal chemist learns is “high cLogP is bad!”. Very rudimentary and simple really. It’s people who are incapable of creating anything of value that are making a career out of banging on this stuff over and over and over and over again.

  22. MedChem says:

    #16, 9, 20
    One of the first things any medicinal chemist learns is “high cLogP is bad!”. Very rudimentary and simple really. It’s people who are incapable of creating anything of value that are making a career out of banging on this stuff over and over and over and over again.

  23. Happy Hanks Auto Surplus, Dallas Texas says:

    TROLL ALERT on #19.

  24. Happy Hanks Auto Surplus, Dallas Texas says:

    Sorry, now he’s #18.

  25. pharmadude says:

    Useless stuff that some office bound loser med chem manager will wave around along with thier copy of lean sigma…hey didn’t that lean sigma crap come out of AZ too? I don’t want to be too harsh on this paper but the authors must realize how useless it is? For all we know this paper is little more than a study in what different companies choose to patent, which is of no value at all. Then you toss in the different targets, then add the error values in clogP, which may well be off by entire units for certain structures, and this whole paper is worthless.

  26. Johny Technoska says:

    Unless we start seeking new targets, small molecule research will be done in a span of 10 years (20-25 for the Process Folks). It is game over.

  27. Franchise says:

    The companies which you did mention are the biggest companies and also well known in world.They are almost equally providing the best services in form of quality medicines.

  28. London Chemist says:

    One of the authors (PL) has been generating papers with similar analysis at the rate of about one a year for most of the last 5 or 6 years. If this is so insightful, then why is the site at which they (the authors) work being closed?

  29. TX Raven says:

    AZ and PFE have the lowest cLogP values, and Wyeth, Lilly, BI and Merck the highest…
    If cLogP was so critical, would you not expect to see that the clinical compounds for the former are failing less than for the latter organizations?

  30. Voltan says:

    Rumour is PL is joining GSK as a consultant – looks like your “problem” is about to be remedied!
    My sympathy….

Comments are closed.