[quick links] Scientific misconduct and scientific inability

• Janet D. Stemwedel has three interesting posts about scientific misconduct
• Comments by David Smith and Andrew Gelman, with follow-ups, about the misuse of statistical tests.

Still no silver bullet for hypothesis testing

There is an interesting post at dechronization about the applicability of one-tailed hypothesis (in a nutshell, the question of when it is appropriate to assume that only very high values are interesting but not very low ones). Andrew Gelman also puts a comment this week on the temptation of overusing p-values.

And then there is this article on ScienceNews claiming that statistical illiteracy is pervasive in Science. The author  claims that many studies have statistical flaws, and for me the message is that not all literature is relevant. We won't find groundbreaking research at every corner, but I would add that this is true for any field... I understand that if this widespread neglect of the statistical theory is systematic within some field, it represents a bigger problem. But in this case their solution of embracing Bayesian methodology will just not work: a Bayesian black box can be more dangerous than a classical one. Because the problem is in the reliance of the black box - a software or a protocol that gives you the illusion of answering your question unattended.

Of course I'm focusing only in one aspect of the article, but I wanted to say that there is no shortcut to improving the quality of scientific studies: authors, reviewers and competitors (when the previous two fail) must get acquainted with the statistical methods (be they classic or Bayesian) behind the tools they use blindly.

ps: HT to Daniel Ferrante for the link.

Communication package between R and Mesquite

Information just received from the R-sig-phylo mailing list:

We'd like to announce the beta release of packages for communication between Mesquite and R. Mesquite is a graphical application for phylogenetic and evolutionary analysis; R is a system for statistical computation primarily oriented around text-based programming. We have enabled communication in both directions:

• RMesquite enables R programs to make use of Mesquite, either as a graphical interface for editing trees and data, or to provide its broad array of phylogenetic calculations to R (http://rmesquite.r-forge.r-project.org).
• Mesquite.R enables Mesquite to make use of functions in R (http://mesquiteproject.org/packages/mesquite.R/MesquiteCallsR/MesquiteR.html). At the moment these include functions in the R packages ape and diversitree.

Only a handful of functions have been enabled in each direction; we seek your input as to what functions would be of most interest to R users and Mesquite users.

You may consider this as an experimental release for thrill-seekers. Please tell us if something doesn't seem to work or is confusing -- feedback is much appreciated. We hope to continue to strengthen the links between R and Mesquite.

Wayne Maddison, UBC
Hilmar Lapp, NESCent

Recent CUDA applications

Taking a look at Nvidia's CUDA zone I found some recent additions - besides important ones that I already knew like the Smith-Waterman and the phylogenetic likelihood algoritihms. Here is a list of these new programs and libraries that take advantage of GPUs:

1. Expectation Maximization algorithm for Gaussian Mixture Models
2. Multilevel algorithm for MDS (multidimensional scaling)
3. Multiclass classifier based on SVMs (support vector machines)
4. GPU computing in the R statistical environment (BLAS lowlevel routines and dist(), hclust(), cor(), and granger.test() functions)
5. Smith-Waterman algorithm for sequence database search
6. Quicksort library
7. Feature finding algorithm for mass spectrometry
8. Smith-Waterman algorithm for sequence alignment
9. MDR (multifactor dimensionality reduction) algorithm for detecting genetic epistatic interactions (white paper)
   

The last reference is just a two-page summary, and the 8th reference is behind a paywall - I couldn't access it myself - but the slides are available here.

Powered by ScribeFire.