Inspired by a discussion with a colleague, I'm going to take one more crack at the recent discussion here about the J. Med. Chem. DHFR paper. Those of you with an interest in the topic, read on. Those whose interest has waned, or who never had much interest to start with, take heart: other topics are coming.
It's clear that many people were disappointed with my take on this paper, and my handling of the whole issue. Let me state again that I mishandled the biology aspects of this one thoroughly, through carelessness, and I definitely owe this apology to the authors of the paper (and the readers of this site) for that.
Of course, that's not the only arguable thing about the way I handled this one. As I spent paragraphs rambling on about in yesterday's post, there's a chemical aspect to the whole issue as well, and that's what caught my eye to start with. I think one of the things that got me into trouble with this one is two different ways of looking at the world. I'll explain what I mean, and you can judge for yourself if I'm making any sense.
The authors of the paper (and its reviewer who commented here) are interested in D67 dihydrofolate reductase, from a biological/enzymological perspective. From this viewpoint - and it's a perfectly tenable one - the important thing is that D67 DHFR is an unusual and important enzyme, a problem in bacterial resistance, interesting in its own right as a protein with an odd binding site, and for all that, still has no known selective inhibitors. Anything that advances the understanding of the enzyme and points toward a useful inhibitor of it is therefore a good thing, and worth publishing in J. Med. Chem., too.
I come in from a different angle. As someone who's done fragment-based drug discovery and takes a professional interest in it, I'll take a look at any new paper using the technique. In this case, I gave the target much too cursory a look, and filed it as "DHFR, bacterial enzyme, soluble, X-ray structures known". In other words, a perfectly reasonable candidate for FBDD as we know it. Once I'd decided that this was a mainstream application of something I already have experience with, I turned my attention to how the fragment work was done. By doing so, I missed out on the significance of the DHFR enzyme, which means, to people in the first camp, that I whiffed on the most important part of the entire thing. I can understand their frustration as I brushed that off like a small detail and went on to what (to them) were secondary matters.
But here's where my view of the world comes in. As a drug discovery guy, when I read a paper in J. Med. Chem., I'd like to see progress in, well, the medicinal chemistry of the topic. That was the thrust of my blog post yesterday: that I found the med-chem parts of the paper uncompelling, and that the application of fragment-based techniques seemed to me to have gone completely off track. (I havne't mentioned the modeling and X-ray aspects of the paper, as Teddy Z did at Practical Fragments, but I also found those parts adding nothing to the worth of the manuscript as a whoel). The most potent compounds in the paper seem, to me, to be the sort that are very unlikely to lead to anything, and are unlikely to show selectivity in a cellular environment. If the paper's starting fragment hits are real (which is not something that's necessarily been proven, as I mentioned in yesterday's post), then it seems to me that everything interesting and useful about them is being thrown away as the paper goes on. From the other point of view, things are basically the opposite - the paper gets better and better as the compounds get more potent.
But here's where, perhaps, the two viewpoints I spoke of earlier might find something in common. If you believe that the important thing is that selective inhibitors of D67 DHFR have finally been discovered, then you should want these to be as potent and selective as possible, and as useful as possible in a variety of assays. This, I think, is what's in danger of being missed. I think that a fragment-based effort should have been able to deliver much more potent chemical matter than these compounds, with less problematic structures, which are more likely to be useful as tools.
I'll finish up by illustrating the different angles as starkly as I can. The authors of this paper have, in one view of the world, completed the first-ever fragment screen against an important enzyme, discovered the first-ever selective inhibitors of it, and have published these results in a prestigious journal: a success by any standard. From my end, if I were to lead a drug discovery team against the same enzyme, I might well see the same fragment hits the authors did, since I know that some of these are in the collections I use. But if I proceeded in the same fashion they did, prosecuting these hit compounds in the same way, I would, to be completely honest about it, face some very harsh questioning. And if I persevered in the same fashion, came up with the same final compounds, and presented them as the results of my team's work, I would run the serious risk of being fired. Different worlds.
Update: Prof. Pelletier sends the following:
I certainly have been following this with interest, and learning much from it – not just science.
Throughout the week, I have appreciated your civil tone – many thanks. I willingly accept your apology, just as I accept the constructive criticism that will improve our future work. I think your ‘two-worlds’ point of view smacks of truth. The bottom line from my point of view is that I’m open to collaboration with a real fragment library: if anyone is interested in making this better, they should contact me. I’d be delighted to work with more than what can be scavenged from neighbouring labs in an academic setting.
Your bloggers’ response to this come-and-go was fascinating: the process was admired to an extent that surprised me. A number of responders point out that there are currently few occurrences of open exchange on these blogs and – sorry to disappoint hard-core bloggers – it does not endear me to the blogging process. I don’t blog because I can’t stand anonymous, frequently disrespectful and sometimes poorly researched comments. I nonetheless hope that this will open the door to a more transparent blogging process in the long run.
For any who care, I am brave, not at all desperate, and definitely a woman. ; )
If you feel any of this would be of interest for your blog, please feel free to post. Thanks for seeing this through rather than shaking it off.