Statistical Optimality in Bioinformatics: Theory vs Practice
Organizer: Rudy Guerra, Rice University
Chair: Rudy Guerra, Rice University
Bioinformatics is in large part defined by the collection, storage and analysis of large biological data sets, which in most cases correspond to some aspect of DNA sequences. The substantive biological or biomedical problems underlying the data are often complex and not amenable to standard statistical models and analysis. As such, computational and statistical algorithms have found much application in bioinformatics. Such algorithms may be founded on basic biological principles, but their statistical behavior is usually limited to knowledge obtained from simulation. This session addresses the role of statistical optimality in bioinformatic problems. More generally, in the face of a complex scientific question, how do statisticians proceed to provide the "best" analysis possible when it may difficult to even cast the problem under a formal statistical framework leading to "optimal" results.