reproducibility

TL;DR We should get science undergraduate students programming by introducing R & Python in all first year science labs, and continuing throughout the undergraduate classes. Why? I’ve previously encountered ideas around getting graduate students to get programming, because to do the analyses that modern science requires you need to be able to at least do some basic scripting, either in a language like Python or R or on the command line.

TL;DR NIH recently introduced a reproducibility initiative, extending to including the “Authentication of Key Resources” page in grant applications from Jan 25, 2016. Seems to be intended for grants involving biological reagents, but we included it in our recent R03 grant developing new data analysis methods. We believe that this type of thing should become common for all grants, not just those that use biological/chemical resources. NIH and Reproducibility There has been a lot of things published recently about the reproducibility crisis in science (see refs).

TL;DR Currently available methods to discover metal geometries make too many assumptions. We were able to discover novel zinc coordination geometries using a less-biased method that makes fewer assumptions. These novel geometries seem to also have specific functionality. This work was recently published under an #openaccess license in Proteins Journal: Yao, S., Flight, R. M., Rouchka, E. C. and Moseley, H. N. B. (2015), A less-biased analysis of metalloproteins reveals novel zinc coordination geometries.

I don’t remember how I got on this, but I believe I had a recent twitter exchange with some persons (or saw it fly by) about pushing R package vignettes to the web after building and checking on travis-ci. Hadley Wickham pointed to using such a scheme to push the web version of his book after each update and the S3 deploy hooks on travis-ci. Deploying your html content to S3 is great, but given the availability of the gh-pages branch on GitHub, I thought it would be neat to work out how to deploy the html output from an R package vignette to the gh-pages branch on GitHub.

Science is built on the whole idea of being able to reproduce results, i.e. if I publish something, it should be possible for someone else to reproduce it, using the description of the methods used in the publication. As biological sciences have become increasingly reliant on computational methods, this has become a bigger and bigger issue, especially as the results of experiments become dependent on independently developed computational code, or use rather sophisticated computer packages that have a variety of settings that can affect output, and multiple versions.

I have been watching the activity in RStudio and knitr for a while, and have even been using Rmd (R markdown) files in my own work as a way to easily provide commentary on an actual dataset analysis. Yihui has proposed writing papers in markdown and posting them to a blog as a way to host a statistics journal, and lots of people are now using knitr as a way to create reproducible blog posts that include code (including yours truly).