ABSTRACT
The reconstruction of phylogenetic (evolutionary) trees from molecular sequence data is a comparatively old problem in bioinformatics, given that Joe Felsenstein’s seminal paper [1] on computing the maximum likelihood score on trees was already published in 1981, that is, almost 3 decades ago. However, signifi cant advances in wet-lab molecular sequencing techniques with the introduction of, for example, the 454 sequencers [2], are generating a highly challenging, unprecedented molecular data fl ood. In addition, recent years have witnessed the emergence of multicore and other parallel architectures such as graphics processing units (GPUs) or the IBM Cell that pose new challenges to the fi eld of phylogenetic or phylogenomic analysis (reconstruction of phylogenies at the genome scale), in particular with respect to orchestrating the phylogenetic likelihood function (PLF). In fact, the phyloinformatics community faces a continuous struggle to keep
up with the rapid speed of data accumulation and provide ever more scalable and powerful analysis tools; that is, we just try to keep pace with data accumulation.
