We have a Nature paper (for a few days, anyway) …

6 04 2015

… well a Nature Methods paper (but still, impact factor > 25!), just submitted today. Now for a nerve-wracking week of checking email every 5 minutes to see if it made the first cut and got sent to referees …. (Nature bounces 90% MSs at this stage; not sure about Nature Methods).

Screen Shot 2015-04-06 at 4.55.48 PM

During the submission process, I realized that I’d written that Jamie had run ‘Monte Carole’ simulations (which perhaps involve three doors and a cat), which I corrected just in time; and I misread an email that said ‘You’ve been registered by Nature Methods‘ as ‘You’ve been rejected by Nature Methods‘ and nearly had a heart attack.

Melodrama aside, what about the science? We showed that you can house three mice of different strains together (C57, Balb/c and DBA) without any welfare problems, and also without messing with their strain-typical phenotypes. Because they look so different, this means that they don’t have to be ear-clipped or otherwise individually marked, but more importantly, the resulting analyses are intrinsically way more powerful than housing mice in same-strain cages, because in this split-plot design, for a given number of cages, the denominator df is bigger than it would be in a conventionThree strain pic 1_Bal design (especially if the study is small; the difference vanishes above about 60 cages).

We also confirmed that mixed-strain housing doesn’t increase data variability, since that would screw up power (in fact it oddly did the opposite, reducing variability thence boosting power yet further — so that for some analyses and some variables, this design could reduce the number of mice needed by half).

Thanks to Laura for the nice pic. And thanks to Jonathan for being a co-author: the first time we’ve worked on anything together.