Labels

Wednesday, April 15, 2015

Normalizing productivity and impact to lab size...

[I've edited the data to include a model small town grocer lab (now the red data point), and to correct the calculations for Figure 4, which were nonsense on the first post]

The following relates to some recent posts over at DM's blog with regard to an essay penned by Andrew D. Hollenbach in ASBMBTODAY. Dr. Hollenbach makes a point in his essay of running a small laboratory, and this got me thinking about productivity as a function of lab size. It's not my intent here to venture an opinion on the pros and cons of productivity metrics &c, but I'd like to post some preliminary figures I rustled up in which I've attempted to graphically present the two main metrics of productivity as a function of laboratory size for a number of heavy hitting labs in neuroscience and cancer (this is a small sample size at this point, and a not remotely randomized one at that, so... pinch of salt &c). I've included a data point based on the available information about Dr. Hollenbach's lab, in addition to a point representing a model "riff raff" lab with one PI, one postdoc, and an annual productivity of one paper (IF ~3.5), which is shown by the red dot.

I know that these metrics have been measured against lab funding before, but I'm not sure I've seen them measured against laboratory labor force directly. Laboratory sizes were ascertained purely based on lab websites and include only grad students, postdocs and postdoc-like research staff (no lab managers, techs, or IT staff*). Clearly, there's no telling what kind of turnover has been going on year by year over the period of analysis, so I'm taking the lazy approach of assuming lab sizes are not changing in absolute number and composition over this time frame (but, again, salty, salty, pinchy, pinchy &c). Only senior/corresponding author publications are counted. 5-year journal Impact Factors were used where possible, with 2013 numbers used when a 5-yr average was not available.

I restricted my analysis to relatively recent productivity over a 3 yr period between the beginning of 2012 and the end of 2014**.

Here it all is...

Figure 1. Displays the relationship between the number of publications per year, for the three years of 2012-2014, and the number of active research personnel in the laboratories sampled. (The red datum represents a model small town grocer outfit of 1 PI and 1 postdoc, with a productivity of 1 paper/yr and an IF average of 3.5)


Figure 2. Mean yearly publication output for the three years of 2012-2014 normalized to the number of active research personnel in the laboratories sampled. (The red datum represents a model small town grocer outfit of 1 PI and 1 postdoc, with a productivity of 1 paper/yr and an IF average of 3.5)



Figure 3. Shows the relationship between lab personnel number and the mean Impact Factor of the journals in which articles were published in the years 2012-2014. (The red datum represents a model small town grocer outfit of 1 PI and 1 postdoc, with a productivity of 1 paper/yr and an IF average of 3.5)

Figure 4. Here I've normalized the average impact for a given year (sum total of journal IFs for all pubs over 3 yrs divided by 3***) to the number of active research personnel in the laboratory. (The red datum represents a model small town grocer outfit of 1 PI and 1 postdoc, with a productivity of 1 paper/yr and an IF average of 3.5)

Figure 5. I thought I'd add in a chart showing the relationship between the impact/lab member/yr and the lab size just for good measure. (The red datum represents a model small town grocer outfit of 1 PI and 1 postdoc, with a productivity of 1 paper/yr and an IF average of 3.5)



* One can argue for the inclusion of these, but for this preliminary analysis at least, I've chosen to exclude them
** Why only 3 yrs, why not 5? - I'll certainly bulk out the numbers in terms of years and laboratories covered later, in the meantime this whole exercise was really just a quick break from end of semester marking.
*** This chart was updated after a facepalm moment over the original cell equation. I think this is more informative.

Saturday, March 28, 2015

Life on the Rusty Edge of Pharmacology II...

Well, dang, we ended up getting a pretty good handle on what The Other is, and how it relates to This and That after all!

So that's nice.



Friday, October 31, 2014

Life on the Rusty Edge of Pharmacology...

So, my working hypothesis was that This happened because of That.
So I applied a drug reported to inhibit That with the prediction that it would consequently inhibit This.
The drug meant to inhibit That actually ended up causing more of This, suggesting that This probably isn't strictly about That after all, and that the That-inhibiting drug is likely doing a little of The Other...
... which opens up a new and exciting avenue of investigation as we move forward to unravel the nature of The Other and its relationship to This and That*.




* Nah, I'm kidding, nobody is going to fund that study. That data is probably going to be buried until it's unearthed millennia from now by aliens picking over the bones and assorted detritus of human civilization. 

Monday, September 29, 2014

Worth the wait...



Apparently this set is pretty popular - which is encouraging in its way - so it was on backorder for a fair while. Of course, the progressive vibe was rocked a little by the fact that the first thing my daughter asked while rummaging through the pieces was why the dark-haired scientist looked so angry. I said it was probably a frown of concentration, but figured it was also worth telling her about Lisa Meitner and Rosalind Franklin, by way of possible explanation for the lack of smiley-facedness in these particular characters. But I don't think she grasped the gravity of the information, and her 6 yr old mind certainly couldn't fathom why missing out on some lame "prize" that didn't come with cake and/or ice cream would be some sort of a big deal. A conversation for another time perhaps.

Monday, August 04, 2014

Tuesday, July 15, 2014

Lab blues and other humbug...

Teaching pharmacology: where one talks about selective drugs with specific effects that tell lovely, tidy stories about how this phenomenon is caused by that mechanism leading to the effective treatment of such-and-such a condition.

Doing pharmacology: where one uses drugs that have multiple cellular targets - many of which are as yet ill-defined/unidentified - with completely different downstream effects depending on the cell type, species, protein expression levels and localization, temperature, stage in the lunar cycle, the NASDAQ index and a few other things besides, and that rarely have a remotely sensible chain of causality linking their cellular effects to the actual clinical outcome.

Wednesday, July 02, 2014

Snap shots: Elite male faculty in the life sciences employ fewer women, J. M. Sheltzer and J. C. Smith, 2014

"For instance, Moss-Racusin et al. (15) sent science faculty identical resumes for a laboratory manager position in which only the name and gender of the applicant were changed. The applicant with the male name was judged to be more competent and hirable and offered a larger starting salary than the female applicant."
From Sheltzer and Smith, PNAS, 2014