https://github.com/Hickerlab/msBayes
Raw File
Tip revision: 7613675e8785be44e5d06e9b78dfef62b2e58add authored by Isaac Overcast on 22 February 2022, 13:01:51 UTC
Merge pull request #4 from ntakebay/master
Tip revision: 7613675
ChangeLog
* Release Version 2010
- Multi-locus simulation
- Uses IM file format as the input file
- batch configuration file has a different format.
- acceptRej.pl has different output.  Use categorical regression
  for discrete variables.

* Release Version 20081106
- msprior.c
 - When Numtau (Psi) is constrained, Tau 1 should always be < Tau2 etc 
   (If Psi constrained to 2, then Tau1 < Tau2 < Tau3 <Tau ).  This was
   not working in previous released version 20080613.  This is fixed.

* Release Version 20080515
- obsSumStats.pl
 -   Output contains a header, easier to identify which column
     correspond to which summary statistics
- msbayes.pl
 -  Psi can be constrained to be a certain value
 -  Shannon Index is added as a new summary statistics
 -  sub-parameters and hyper-parameters can be constrained
 -  Reduced File I/O to improve the performance when the program is
    running over network file system.
 -  During the simulation run, results accumulate in temporary files 
    (usually in /tmp), and the results get copied over and reformated 
    to the final output after all simulations are done.
- acceptRej.pl
 -  Bayes factor plots are produced
 -  acceptRej.pl can handle much larger number of simulation.  The number of
    simulation used to be limited by the memory.  Now initial
    accept-rejection part is done by C program before R start to do analysis.
	
* Mon Jul 23 2007 Naoki Takebayashi <ffnt@uaf.edu>
- Detailed logs are now kept in svn
- This file may be used to record major changes between releases.

* Mon Nov  6 2006 Naoki Takebayashi <ffnt@uaf.edu>
- obsSumStats.pl was using the old format (10 or 11 column formats).
  We removed one of the two Ts/Tv ratio column and gamma rate columns, so
  new input files contains 8 or 9 columns
- in msprior.c, upper limit for ancestral population size was incorrectly
  gParam.upperAncPopSize*gParam.upperTheta*gParam.upperTheta/theta.
  Correct upper limit is gParam.upperAncPopSize*gParam.upperTheta/theta
- README was updated to explain upper ancestral population size better.
- Makefile updated, so one can make static binary with "make STATIC_LINK=YES".
  For Mac OS-X with fink, this cause static linking to gsl, but dynamic link
  to other system libs.

* notes for a version before 20061005 Naoki Takebayashi <ffnt@uaf.edu>
- Modified msprior.[ch] -- gParam.upperTheta (was 40.5) and
  gParam.lowerTheta (was 0.5) is used.
- Modified setup.c
- Modified msprior, so the first element of output line is now upper theta,
  which is required to stat program.
- modified samplesDQH_rej_16tx.c
  Takes option flags, which specify upper bound to theta (-T).
  nadv used to be an argument, but it is changed to option (-a).
- modified msbayes.pl to pass upper theta around

back to top