Statistical Alignment MCMC Workbench
Since about year 2000 there has been a shift in sequence analysis methods from optimisation based methods to methods based on underlying stochastic models of evolution of sequences. This is a very positive trend that allows for much better and reliable analysis of sequences and genomes. However, statistical alignment is still a computationally very hard problem and it is important to devise efficient algorithms. It seems the way ahead is the use of MCMC algorithms and the aim of this project is to assess the efficiency of different MCMC statistical alignment methods on simple data sets to assess their computational properties. All MCMC methods for integrating over all possible alignments should be reviewed and implemented. There are probably 4-6 methods. This does not involve any investigation of different phylogenies.
Printer Friendly Version 