-
Notifications
You must be signed in to change notification settings - Fork 1
Performance testing / improvement on migrate-n
mipearson/migrate-n
Folders and files
Name | Name | Last commit message | Last commit date | |
---|---|---|---|---|
Repository files navigation
------------ Migrate 3.x ------------ 3.2.7 released January 12 2011 Quick installation guide ------------------------ - Unpack the compressed distribution file - Macos, Unix: double click the file: install_migrate Source files: configure; make or make install_migrateshell or make mpis or .... Content ------- - Overview - Analyses summary - Computer systems - History of MIGRATE - Distribution - Installation - Documentation - Example folder - Troubleshoothing - Disclaimer Migrate estimates population parameters, effective population sizes and migration rates of n populations, using genetic data. It uses a coalescent theory approach taking into account history of mutations and uncertainty of the genealogy. The estimates of the parameter values are achieved by either a Maximum likelihood (ML_ approach or Bayesian inference (BI). The output is presented in an TEXT file and in a PDF file. The PDF file eventually will contain all possible analyses including histograms of posterior distributions. Currently only main tables (ML + BI), profile likelihood tables (ML), percentiles tables (ML), and posterior histograms (BI) are supported in the PDF. Currently the following data types are supported: -DNA sequence data - finite sites model: F84, Kimura two-parameter - finite sites model + rate variation among sites: F84 + Gamma, F84 + arbitrary rates -SNP data (single nucleotide polymorphism) - SNP are derived from sampled sequences and are completely linked except that we know that the sites are variable [not thoroughly tested] -Microsatellite data - stepwise mutation model - brownian motion model: a continuous approximation to the stepwise mutation model. the approximation breaks down when the population sizes are small (Theta<5) -Electrophoretic marker data (infinite allele model). Analyses (+ = new in this version) ---------------------------------- - Estimation of population sizes and migration rates of a migration matrix model, or arbitrarily subsets of a migration matrix model, or an n-island model. Allowing for a geographic distance matrix so that effects geographic can be removed out of the analysis.(ML + BI) - Profile likelihood curves deliver approximate confidence intervals.(ML) - Allows approximate likelihood ratio tests and model selection using Akaike's Information criterion. (ML) + Marginal likelihood calculation to assist calculation of Bayes Factor (BI) - Plots of overall immigration and emigration per population. (ML) + Allows a variable mutation rate AMONG loci. (ML [+BI]) - For sequences: allows a variable substitution rate among sites. (ML+BI) + For microsatellites: allows the definition of repeatnumber and use of fragment length as input - Facilitates analyses of multimodal search space distributions with heating scheme and/or multi-run analyses. (ML+BI) - Histogram of migration events over time (ML + BI) + Plot of expected parameters through time (skyline plots) for all sizes and migration rates. + Dated samples (ML + BI) + Relabeling and mergin of populations + Random subset of individuals per population Computer systems ---------------- You can fetch Migrate from the website http://popgen.sc.fsu.edu as source code or binary executables. Currently I supply - Macintosh - Macintosh G4 (PPC): migrate-n - Macintosh Intel: migrate-n migrateshell.app (includes migrate-n and fastmigrate-n - Windows: migrate-n.exe - Linux - 32bit: migrate-n - 64bit Opteron: migrate-n The file is compressed as tar.gz or as zip file. The documentation contains information about how to compile and use a parallelized version of migrate so that it can run concurrently on computer clusters (using MPI [preferrably OpenMPI]). History about bug fixes and new features ---------------------------------------- read the HISTORY file. Distribution ------------ Migrate can be fetched from the www-site http://popgen.sc.fsu.edu/ Installation ------------ (a) Binaries Unpack, download and read the documentation and try the program on a small data set. For UNIX systems the binary can go to standard directories (e.g. /usr/local/bin), the rudimentary man page can go to the /usr/local/man/man1. (b) Source (UNIX) 1. gunzip -c migrate-3.1.2.src.tar.gz | tar xf - or use tar xvfz migrate-3.1.2.src.tar.gz [this creates a directory "migrate-3.1.2" with subdirectories "src", and "examples" in it.] 2. cd migrate-3.1.2/src 3. type "./configure" This will create the Makefile [on some machines, you can squeeze out a little more speed by trying to use (if you use the bash shell:) CC=cc ./configure or (if you use a csh shell type:) env CC=cc ./configure; or use the Intel compiler icc on Linux machines] 4. type "make" (please report warnings and especially errors). If you have a multiprocessor machine you perhaps want to try "make thread" (this allows parallel execution of chains when using the heating scheme). "make" will produce migrate with PDF support, this might not work on all machines, if it fails please tell me. The result of the compilation should be an executable "migrate-n" in the src directory [it is called "migrate-n" because on some computer system there is a system program called "migrate"] 5. make install This will install the programs and man-page into usr/local/bin, /usr/local/man/man1 [you need to be root to do this; this step is not necessary, to use the program, but it would be convenient for all users of your system] or move migrate-n to $HOME/bin or some other convenient place. 6. change directory to example run "migrate-n parmfile.testml" and "migrate-n parmfile.testbayes", on my 2008 macosx laptop the BI run took about 2 minutes. If these two tests fail, please let me know! Documentation ------------- You need to download it separately from http://popgen.scs.fsu.edu/Downloads.html It is a PDF file and called migratedoc.pdf. The pdf file can be viewed and printed using Acrobat or any other PDF viewer. Examples -------- In the directory "example" you can find some example data sets. You might wan to try the two parmfiles.testml and parmfile.testbayes Use the Terminal.app (on macosx), or xterm (Linux), or cmd (Windows), change directory to the example directory and then execute for mac and unix: ../migrate-n parmfile.testbayes for windows: ..\migrate parmfile.testbayes or the ml version. Disclaimers ----------- This software is distributed free of charge for non-commercial use and is copyrighted. Of course, we do not guarantee that the software works and are not responsible for any damage you may cause or have. Copyright --------- (c) 1997-2003, Peter Beerli and Joseph Felsenstein, Seattle. (c) 2004-2011, Peter Beerli, Tallahassee Fan-mail, complaints, questions and error-reports ------------------------------------------------- Peter Beerli beerli@fsu.edu Last update: December 6 2010
About
Performance testing / improvement on migrate-n
Resources
Stars
Watchers
Forks
Releases
No releases published
Packages 0
No packages published