SiPhy implements rigorous statistical tests to detect bases under selection from a multiple alignment data. It takes full advantage of deeply sequenced phylogenies to estimate both unlikely substitution patterns as well as slowdowns or accelerations in mutation rates.
It can be applied as an Hidden Markov Model (HMM), in sliding windows, or to specific regions. For more detailed information see documentation.
A quick tutorial to get you up and running.
To cite your use of SiPhy, please reference our publication Identifying novel constrained elements by exploiting biased substitution patterns, Bioinformatics 2009 25(12):i54-i62.