Mcmcphase/Mlphase

From Vd

Jump to: navigation, search

PHASE: a software package for PHylogenetics And Sequence Evolution

Contents

General description

PHASE is a package for performing molecular phylogenetic inference using Bayesian or Maximum Likelihood inference. The package is designed specifically for use with RNA sequences that have a conserved secondary structure, e.g. rRNA and tRNA, but it also handles noncoding DNA, protein coding DNA, amino acid sequences or mixed data sets.

The PHASE package consists of two main programs:

  • mlphase performs maximum-likelihood inference with (limited) tree-search capabilities
  • mcmcphase is a Bayesian phylogenetic inference program.

There are nine other smaller programs in the package:

  • optimizer is a smaller version of mlphase designed to optimize user-defined trees
  • distphase commputes pairwise-ML distance estimates between species
  • mcmcsummarize is used with mcmcphase to output and summarize the results of a MCMC run
  • simulate generates sequences according to a specified evolutionary model
  • analyzer checks the content of the molecular sequences
  • likelihood computes the likelihood of the data given an evolutionary model (tree + substitution model)
  • extractfromDCSE is a crude program for parsing a DCSE sequence alignment into an alignment file where structures and nucleotides are dissociated.
  • secondarystructconsensus is a crude program for producing a consensus structure from the aligned secondary structures of a set of species
  • splitdataSet splits a given datafile into smaller parts

Application homepage, including an Online manual.

Using PHASE

The main programs take as input a data file in a slightly modified PHYLIP format, and a control file containing settings for the analysis.

Data File

The data file consists of 4 parts:

1) A line containing the number of taxa, the number of characters and the type of data (DNA, PROT, CODON or STRUCT)

2) An optional line containing a secondary structure mask in dot-parenthese notation

3) The data matrix. Taxon labels must be without blanks and be separated from the data by one or more blanks

4) An optional line containing a class designation for each site in the data matrix

Different categories of nucleotides/amino acids can be designated to different classes and subjected to different substitution models in the analysis. An obvious choice is to designate paired stem-regions and unpaired loop-regions in RNA sequences to separate classes, but separate classes can also be used for different genes or when combining different types of sequences. In the example below, an RNA sequence with two hairpins and a protein coding sequence with 3 codons have been concatenated. loops, stems and protein coding sites have been designated classes 1, 2 and 3 respectively.

Example: 

3 28 STRUCT
          ((....)).((((..))))123123123
Species1 UUAUGGAACACGUUAACGUGGTAACCCA
Species2 UUAUGGGACACUUUAAAGUGGAAACCCT
Species3 UUAUGGGACACGUUAACGUGGTAATCCT
          2 2 1 1 1 1 2 2 1 2 2 2 2 1 1 2 2 2 2 3 3 3 3 3 3 3 3 3

Control File

The control file varies somewhat from program to program. See the manual on the Phase homepage for details. The file is organized in blocks of data, each starting with {BLOCKNAME} and terminating with {\BLOCKNAME}.

  • The {DATAFILE} block contains a link to the data file and some descriptive parameters
  • The {MODEL} block defines a substitution model for each of your classes
  • The {TREE} block contains information about your preferred tree type
  • The {PERTURBATION} block contains two sub-blocks, {PERTURBATION_TREE} and {PERTURBATION_MODEL}, containing the proposals and priors used for bayesian inferences

The file terminates with some general run settings

Example of an mcmcphase control file: 

{DATAFILE}
   Data file = C:\MyData\datafile.txt
   Interleaved data file = no
   Interleaved structure = no
   Heterogeneous data models = yes
{\DATAFILE}

#Set up a MIXED model with REV for loops, 7D for stems and YNH98 for codons
{MODEL}
   Model = MIXED
   Number of models = 3

   #Unpaired loop regions
   {MODEL1}
     Model = REV
     Discrete gamma distribution of rates = yes
     Number of gamma categories = 6
     Invariant sites             = yes
   {\MODEL1}

   #Paired stem regions
   {MODEL2}
     Model = RNA7D
     Discrete gamma distribution of rates = yes
     Number of gamma categories = 6
     Invariant sites             = yes
   {\MODEL2}

   #Codons
   {MODEL3}
     Model = YNH98
     Genetic code file = input-data\GeneticCode\InvertebrateMitochondrial.txt
   {\MODEL3}
{\MODEL}

{TREE}
   Tree = Unrooted MCMC tree
   Outgroup = H_habilis_01
{\TREE}

#Tuning parameters for the MCMC runs.
{PERTURBATION}
   #relative proposals probabilities between the tree and the substitution model
   Tree, proposal priority = 8
   Model, proposal priority = 1

   {PERTURBATION_TREE}
     Topology changes, proposal priority = 10
     Branch lengths, proposal priority = 40
     Hyperpriors, proposal priority = 1
     Branch lengths, prior = exponential(uniform(0,10))
     Branch lengths exponential hyperparameter, proposal priority = 1
   {\PERTURBATION_TREE}

   {PERTURBATION_MODEL}
     Model 1, proposal priority = 10
     Model 2, proposal priority = 10
     Model 3, proposal priority = 10
     Average rates, proposal priority = 1

     {PERTURBATION_MODEL1}
       Frequencies, proposal priority = 2
       Rate ratios, proposal priority = 1
       Gamma parameter, proposal priority = 1
       Invariant parameter, proposal priority = 1
     {\PERTURBATION_MODEL1}

     {PERTURBATION_MODEL2}
       Frequencies, proposal priority = 2
       Rate ratios, proposal priority = 1
       Gamma parameter, proposal priority = 1
       Invariant parameter, proposal priority = 1
     {\PERTURBATION_MODEL2}

     {PERTURBATION_MODEL3}
       Frequencies, proposal priority = 2
       Rate ratios, proposal priority = 1
       Gamma parameter, proposal priority = 1
       Invariant parameter, proposal priority = 1
     {\PERTURBATION_MODEL3}

   {\PERTURBATION_MODEL}

{\PERTURBATION}

Random seed = 11
Burnin iterations = 75000
Sampling iterations = 150000
Sampling period = 15
Output file   = C:\MyData\output
Output format = phylip

PHASE on Titan

The distributed input-data directory is installed as /site/mcmcphase/input-data on titan.

Running PHASE on the Bioportal

The application on the Bioportal for running the PHASE programs is called mcmcphase (there is a different application called phase). It is located in the bio category.

The application can be used to run mcmcphase, mlphase, optimizer, distphase, simulate, and likelihood. When running mcmcphase, one can postprocess the output with mcmcsummarize.

To use the application, upload a suitable control file, as well as the needed data files. In the Application settings, choose which file is the control file, which program to run, and (for mcmcphase) whether to post process with mcmcsummarize.

When the program is run, all files are located in the current working directory, so all files should be specified without any path in the control file, for instance: 

Output file = myOutputFile

(The exception is files in the distributed input-data directory, which is installed as /site/mcmcphase/input-data.

Retrieved from "http://heorogar.titan.uio.no/index.php/Mcmcphase/Mlphase"
Views
Personal tools