December 7, 2018

There was recently a little twitter thread on the obligations scientists have to work on finishing up projects, particularly once the formal association with that project has expired. This made me recall some thoughts I had from the PI perspective on finishing projects one has been funded to work on. So I was primed when these thoughts occurred today.

Once upon a time I ran across an interesting RFA from the NIH that seemed highly targeted for one particular lab. Oh, don’t get me wrong, many of them do seem this way. But this one was particularly…..specific. The funding was generous, we’re talking BigMech territory here. And while I could theoretically pull together a credible proposal with the right collaborators, I really wasn’t the right person. Meh, it was a time of transition in my lab and in my science and I didn’t have anything else to write at the time so….I put a proposal in.

And did not get triaged! whoohoo! Really, I should have been triaged. Clearly the field of proposals was really, really weak. As in not even credible, I am assuming. You know my mantra….always submit something that hits the minimal standards for a reviewer to get behind it. So mine was credible. I just didn’t really have the specific chops on that area of work to support a good score.

The top-suspect lab got the grant, of course.

Which again, was very specific as to the topic and goals. Most interestingly it was for “Approach X” to the general idea when this laboratory had only vaguely flirted with X in their earliest going and had settled upon Approach Y.

I kept a bit of a weather eye on what they were doing with the funds, as you can imagine. After all I was interested in the topic enough to put in a grant proposal.

The Approach Y papers kept on coming. Nary an Approach X paper to be seen. I saw one very limited poster at a meeting once in the middle of their funding interval. And then at the end those data were published, with not much ground traveled. I did mention it was a BigMech, right?

That was a fair bit ago and apparently the laboratory got a second BigMech, a different one, but again clearly devoted to Approach X on the question at hand. And it was a few years in by the time I noticed.

I hadn’t paid the strictest attention but at one point I got a manuscript to review from the laboratory. “Aha”, I thinks to meself, “this is going to really kick ass after so much funding and so few papers”. Both BigMechs are acknowledged, so we’re on solid ground assuming what paid for this work.

Of course the manuscript was underwhelming. Not that anyone else is doing amazingly better solving this particular question. But the point is that they had copious funding over at least seven years and have barely published anything relevant to Approach X take on the topic. This dataset is relevant but far from earthshaking. It just retreads territory published five years before it appeared. And the functional outcome of this work is no better than one-off papers from at least three other groups who have not enjoyed such amazing levels of funding.

Now, the laboratory has published other work. Just mostly on Approach Y which was not the goal according to the original FOA and going by the abstracts on RePORTER.

It’s possible that the group has been working away like beavers on the topic and just never had any positive results. It is not totally foreign to see a laboratory that will only publish big positive hits. But I do not see that in their overall history. There does not seem to be a reluctance to squeak out small papers with one thin asterix showing positive effects. So I have to conclude they just aren’t really working on Approach X much at all.

In a way this makes me angry. I still feel like the bargain we make is to take a serious stab at working on what the grant says. Yeah, yeah a grant is not a contract but in this case there were fairly focused calls for applications. Someone who was actually planning to do the work was not funded. Perhaps more than one someone.

[record scratch]

Wait a minnit dude, doesn’t this apply to you!???!!!!


It has.

I mean, not in the sense of highly targeted original FOA but sure, I’ve had my intervals of what looks like pretty good funding on a topic and have batted below expectations on publications. Of course I know how hard we worked to get something going on the project as proposed. I know that we did the science as best we could. I know what headaches interfered and I know what data exist that I mean to publish* one day. I have certainly credited that project with other somewhat-related side project work to the extent it was kosher.

But a critic could go through this same thought process when they get my manuscript to review.

“That’s IT????!!!!!”

And if they have been less fortunate than I in the NIH grant game, the odds are they are going to rant about how unfair everything is and how I must have sweet political connections and the Systemz is Brokenz.

They aren’t wrong.

And they aren’t exactly right, either.

These scenarios play out all the time in NIH extramural funding land. I don’t know how you do real science and not come up blank once in awhile. Occasionally this is going to play out to the tune of an entire BigMech, right? Maybe?

Otherwise we are only doing utterly safe and conservative science. Is that correct?

*This is where it intersects with the twitter discussion. Writing up old projects, exciting or not, may be hindered if the staff who worked on it have now departed. I’m not going to sidetrack but obviously the obligation to render published science for grant dollars awarded falls heavily on the PI and less so on the trainees. But still, if you have been paid to work on the project, there is an obligation to produce on that project.