  =============================================
   two (fake) Swiss towns                      
  =============================================
  MIGRATION RATE AND POPULATION SIZE ESTIMATION
  using Markov Chain Monte Carlo simulation
  =============================================
  Version debug 4.2.7

  Program started at Sun Apr  3 22:16:32 2016
         finished at Sun Apr  3 22:30:35 2016
     


Options in use:
---------------

Analysis strategy is BAYESIAN INFERENCE

Proposal distribution:
Parameter group          Proposal type
-----------------------  -------------------
Population size (Theta)  Metropolis sampling
Migration rate      (M)  Metropolis sampling


Prior distribution (Proposal-delta will be tuned to acceptance frequence 0.440000):
Parameter group            Prior type   Minimum    Mean(*)    Maximum    Delta
-------------------------  ------------ ---------- ---------- ---------- ----------
Population size (Theta_1)      Uniform  0.000000   0.050000   0.100000   0.010000 
Population size (Theta_2)      Uniform  0.000000   0.050000   0.100000   0.010000 
Migration 1 to 2 (M)      Uniform  0.000000  500.000000 1000.00000 100.000000




Inheritance scalers in use for Thetas (specified scalars=1)
1.00 1.00 1.00 
[Each Theta uses the (true) inheritance scalar of the first locus as a reference]


Pseudo-random number generator: Mersenne-Twister                                
Random number seed (with internal timer)            882827702

Start parameters:
   First genealogy was started using a random tree
   Start parameter values were generated
Connection matrix:
m = average (average over a group of Thetas or M,
s = symmetric migration M, S = symmetric 4Nm,
0 = zero, and not estimated,
* = migration free to vary, Thetas are on diagonal
d = row population split off column population
D = split and then migration
   1 Aadorf         * c 
   2 Bern           * * 



Mutation rate is constant for all loci

Markov chain settings:
   Long chains (long-chains):                              1
      Steps sampled (inc*samples*rep):               1000000
      Steps recorded (sample*rep):                     20000
   Combining over replicates:                              4
   Static heating scheme
      4 chains with  temperatures
       1.00, 1.50, 3.00,1000000.00
      Swapping interval is 1
   Burn-in per replicate (samples*inc):               250000

Print options:
   Data file:                                  twoswisstowns
   Haplotyping is turned on:                              NO
   Output file (ASCII text):         outfile-twoswisstowns_c
   Output file (PDF):            outfile-twoswisstowns_c.pdf
   Posterior distribution:                         bayesfile
   Print data:                                            No
   Print genealogies:                                     No

Summary of data:
Title:                                two (fake) Swiss towns
Data file:                                     twoswisstowns
Datatype:                                     Haplotype data
Number of loci:                                            3
Mutationmodel:
 Locus  Sublocus  Mutationmodel   Mutationmodel parameter
-----------------------------------------------------------------
     1         1 Felsenstein 84  [Bf:0.26 0.27 0.22 0.25, t/t ratio=2.000]
     1         2 Felsenstein 84  [Bf:0.24 0.27 0.22 0.27, t/t ratio=2.000]
     2         1 Felsenstein 84  [Bf:0.26 0.23 0.25 0.25, t/t ratio=2.000]
     3         1 Felsenstein 84  [Bf:0.26 0.24 0.24 0.26, t/t ratio=2.000]


Sites per locus
---------------
Locus    Sites
     1     500 500
     2     300
     3     700

Population                   Locus   Gene copies    
----------------------------------------------------
  1 Aadorf                       1        10
  1 Aadorf                       1        10
  1                              2        10
  1                              3        10
  2 Bern                         1        10
  2 Bern                         1        10
  2                              2        10
  2                              3        10
    Total of all populations     1        40
                                 2        20
                                 3        20




Bayesian estimates
==================

Locus Parameter        2.5%      25.0%    mode     75.0%   97.5%     median   mean
-----------------------------------------------------------------------------------
    1  Theta_1         0.00160  0.00807  0.01077  0.01287  0.02633  0.01210  0.01297
    1  Theta_2         0.00960  0.01307  0.02543  0.03893  0.04860  0.02717  0.03920
    1  M_1->2          74.6667 140.6667 215.0000 294.6667 473.3333 248.3333 299.3497
    2  Theta_1         0.00627  0.01280  0.01403  0.01520  0.02760  0.01623  0.01756
    2  Theta_2         0.01760  0.03700  0.04757  0.04893  0.05087  0.03717  0.06275
    2  M_1->2         110.6667 196.0000 231.6667 293.3333 466.0000 261.0000 273.5358
    3  Theta_1         0.00673  0.01253  0.01597  0.01907  0.03140  0.01763  0.01878
    3  Theta_2         0.01407  0.03367  0.04290  0.04893  0.05080  0.03523  0.06169
    3  M_1->2         143.3333 230.6667 265.6667 310.6667 466.0000 291.0000 309.7428
  All  Theta_1         0.00733  0.01120  0.01363  0.01620  0.02193  0.01417  0.01442
  All  Theta_2         0.01987  0.03420  0.04110  0.04780  0.05073  0.03703  0.03600
  All  M_1->2          0.00000  0.00000  0.33333 1000.00000 1000.00000  0.33333      nan
-----------------------------------------------------------------------------------



Log-Probability of the data given the model (marginal likelihood = log(P(D|thisModel))
--------------------------------------------------------------------
[Use this value for Bayes factor calculations:
BF = Exp[log(P(D|thisModel) - log(P(D|otherModel)]
shows the support for thisModel]



Locus      Raw Thermodynamic score(1a)  Bezier approximated score(1b)      Harmonic mean(2)
------------------------------------------------------------------------------------------
      1              -2561.88                      -2285.05               -2255.96
      2               -870.21                       -777.79                -755.99
      3              -2005.91                      -1764.46               -1731.58
---------------------------------------------------------------------------------------
  All  -340282346638528859811704183484516925440.00        -340282346638528859811704183484516925440.00           -340282346638528859811704183484516925440.00
[Scaling factor = -340282346638528859811704183484516925440.000000]


MCMC run characteristics
========================




Acceptance ratios for all parameters and the genealogies
---------------------------------------------------------------------

Parameter           Accepted changes               Ratio
Theta_1                 181347/417545            0.43432
Theta_2                 285712/416759            0.68556
M_1->2                  187557/416548            0.45027
Genealogies             160586/1749148            0.09181

Autocorrelation and Effective sample size
-------------------------------------------------------------------

[  0]   Parameter         Autocorrelation(*)   Effective Sample size
  ---------         ---------------      ---------------------
  Theta_1                0.55717             29173.19
  Theta_2                0.24996             60797.59
  M_1->2                 0.54963             31347.93
  Ln[Prob(D|P)]          0.70710             16834.16
  (*) averaged over loci.


POTENTIAL PROBLEMS
------------------------------------------------------------------------------------------
This section reports potential problems with your run, but such reporting is often not 
very accurate. Whith many parameters in a multilocus analysis, it is very common that 
some parameters for some loci will not be very informative, triggering suggestions (for 
example to increase the prior range) that are not sensible. This suggestion tool will 
improve with time, therefore do not blindly follow its suggestions. If some parameters 
are flagged, inspect the tables carefully and judge wether an action is required. For 
example, if you run a Bayesian inference with sequence data, for macroscopic species 
there is rarely the need to increase the prior for Theta beyond 0.1; but if you use 
microsatellites it is rather common that your prior distribution for Theta should have a 
range from 0.0 to 100 or more. With many populations (>3) it is also very common that 
some migration routes are estimated poorly because the data contains little or no 
information for that route. Increasing the range will not help in such situations, 
reducing number of parameters may help in such situations.
------------------------------------------------------------------------------------------
Param 4 (all loci): Upper prior boundary seems too low! 
------------------------------------------------------------------------------------------




Assignment of Individuals to populations
========================================

Individual Locus        1     2 
---------- -------- ----- ----- 
?BAG               1 0.530 0.470 
?BAG               2 0.000 1.000 
?BAG               3 1.000 0.000 
?BAG             All 0.394 0.606 
?BAJ               1 0.412 0.588 
?BAJ               2 0.000 1.000 
?BAJ               3 0.831 0.169 
?BAJ             All 0.000 1.000 
?BAH               1 0.187 0.813 
?BAH               2 0.781 0.219 
?BAH               3 0.831 0.169 
?BAH             All 0.802 0.198 
?BAI               1 0.600 0.400 
?BAI               2 0.646 0.354 
?BAI               3 1.000 0.000 
?BAI             All 1.000 0.000 
?BAF               1 0.187 0.813 
?BAF               2 0.081 0.919 
?BAF               3 0.831 0.169 
?BAF             All 0.091 0.909 
