################################################################################
# Parmfile for Migrate debug 2.1.9-x [do not remove these first TWO lines]
# generated automatically on
# Tue Dec 12 15:35:08 2006
#
# please report problems to Peter Beerli
#  email: beerli@scs.fsu.edu
#  http://popgen.csit.fsu.edu/migrate.html
################################################################################
#
################################################################################
# General options
################################################################################
#
# Interactive or batch job usage
#   Syntax: menu= < YES | NO > 
# For batch runs it needs to be set to NO
menu=YES
#
# Specification of length of names of indiviudals
#    Syntax: nmlength=<INTEGER between 0 .. 30>
nmlength=10
#
#
################################################################################
# Data options
################################################################################
#
# Several different main datatypes are possible:
# INFINITE ALLELE: usable for electrophoretic markers,
#                  other markers with unknown mutation model
# STEPWISE MUTATION: usable for microsatellite data or
#                  other markers with stepwise change
#                  from one allele to another
# FINITE SITES MUTATION: standard DNA/RNA sequence mutation
#                  model, usable for DNA or RNA contiguous
#                  sequences or varialbe sites only (SNP)
# GENEALOGY SUMMARY: reanalyzing an old migrate run
#
#-------------------------------------------------------------------------------
# INFINITE ALLELE
#  Syntax: datatype=ALLELICDATA 
#          include-unknown=<YES | NO> with YES unknown alleles
#                are included into analysis, NO is the default
#
#-------------------------------------------------------------------------------
#
# STEPWISE MUTATION
#  Syntax: datatype=<MICROSATELLITEDATA | BROWNIANDATA
#                MICRO specifies the standard stepwise mutation
#                model, the BROWNIAN is an approximation to this
#          micro-threshold=<INTEGER> Default is 10 [MICRO only, NEEDS TO BE EVEN!],
#                smaller values speed up analysis, but might also
#                crash, large values slow down analysis considerably.
#                Change this value only when you suspect that your
#                data has huge gaps in repeat length.
#          include-unknown=<YES | NO> with YES unknown alleles
#                are included into analysis, NO is the default
#
#-------------------------------------------------------------------------------
#
# FINITE SITES MUTATION
#  Syntax: datatype=<SEQUENCEDATA | NUCLEOTIDE | UNLINKEDSNPS | ANCESTRAL
#         SEQENCEDATA: typical linked stretches of DNA, for example mtDNA
#         NUCLEOTIDE: linked DNA stretches, all invariable sites removed
#         UNLINKEDSNPS: each variable site is a locus, DO NOT USE THIS YET
#         ANCESTRAL: instead taking into account all posible states, use
#                use only the most likely state probability, DON'T USE THIS YET
#
#          freqs-from-data=<YES | NO: freq(A), freq(C), freq(G), freq(T)>
#                calculate the prior base frequencies from the data,
#                or specify the frequencies
#          ttratio=<RATIO1 RATIO2 ....> Default is 2.0,
#                ratio between transitions and transversions.
#          seq-error=<VALUE> Default is 0.0, typical values for ABI 3700 
#                sequencers after base calling are around 0.001 (1/650)
#          categories=<VALUE:CATFILE> The categories are integers or letters
#                specified in file called CATFILE, this assumes that all
#                sites belong to known categories, this can be used to
#                weight third positions etc.
#          rates=<VALUE1 VALUE2 ...> the rates are specified arbitrarily or
#                then are from a Gamma distribution with alpha=x, currently
#                 the alpha value gets lost and is not recorded in the parmfile
#          prob-rates=<RATE2 RATE1 ... > These rates can be arbitrary or 
#                generated with gamma-deviated rates and then are derived
#                using Laguerre's quadrature, this should get better
#                results than equal probability methods.
#          autocorrelation=<NO | YES:VALUE> Default is NO
#                autocorrelation makes only sense with rates,
#                VALUE should be >1.0
#          weights=<NO | YES:WEIGHTFILE> The weights are specified
#                in file called WEIGHTFILE, this assumes that all sites
#                belong to known weights, this can be used to weight
#                portions of the sequence etc.
#          interleaved=<YES | NO> Use either an interleaved or 
#                non-interleaved format. Default is NO,
#                interleaved=YES is discouraged
#          fast-likelihood=<YES | NO> Default is YES, use NO when you
#                have many hundred individuals and get strange errors
#                during a run, NO is scaling the conditional likelihood
#                so that very small values are >0.00000
#          usertree=<NO | TREE:TREEFILE | DISTANCE:DISTFILE | RANDOM>
#                Default is RANDOM, NO delivers a UPGMA tree using the data
#                with TREE and DISTANCE the user needs to 
#                give a usertreefile or a pairwise distance file, with RANDOM
#                a random tree will be the starting tree
#
#-------------------------------------------------------------------------------
#
#
datatype=SequenceData
ttratio=2.000000 
freqs-from-data=YES
seqerror-rate=0.000000
categories=1 #no categories file specified
rates=1: 1.000000 
prob-rates=1: 1.000000 
autocorrelation=NO
weights=NO
interleaved=NO
fast-likelihood=YES
usertree=RANDOMTREE
#
################################################################################
# Input options
################################################################################
#
# input file location
#   Syntax infile=FILEPATH
infile=infile.dna
#
# Random number seed specification
#   Syntax random-seed=<AUTO | OWN:< seedfile | value >
#      AUTO           uses computer system clock to generate seed
#      OWN:seedfile   uses file seedfile with random number seed
#      OWN:value      uses number value for seed
random-seed=OWN:285548372
#
# Specify the title of the run, will be overridden by title in datafile
#    Syntax: title=title text [up to 80 characters]
title=infinite alleles geo + custom
#
#
################################################################################
# Output options
################################################################################
#
# Progress report to the window where the program was started
#    Syntax: progress=<NO | YES | VERBOSE>
#          NO       nothing is printed to the console
#          YES      some messages about progress are reported [default]
#          VERBOSE  more messages are reported to console
progress=YES
#
#-------------------------------------------------------------------------------
#
# Recording messages to screen into logfile
#   Syntax logfile=<NO | YES:logfilename>
#       NONE     no recording of progress
#       logfilename  path to logfile
logfile=YES:logfile
#
#-------------------------------------------------------------------------------
#
# Print the data as read into the program
#   Syntax print-data=<NO | YES>
print-data=NO
#
#-------------------------------------------------------------------------------
#
# Print output to file [default is outfile]
#   Syntax outfile=outfilename
outfile=outfile-dna
#
#-------------------------------------------------------------------------------
#
# Print output to a PDF file [default is outfile.pdf]
#   Syntax pdf-outfile=outfilename.pdf
pdf-outfile=outfile-dna.pdf
#
#-------------------------------------------------------------------------------
#
# Report M (=migration rate/mutation rate) instead of 4Nm or 2 Nm or Nm
#   Syntax use-M=<NO | YES> Default is YES, the name 4Nm is ambiguous
#      for non-diploid data
use-M=YES
#
#-------------------------------------------------------------------------------
#
# Plotting parameters: migration versus population size, such that Theta1 x immigration_.1
# this shows the sum of all imigrations int a population
#   Syntax plot=<NO | YES:<BOTH | OUTFILE>:<LOG | STD>
#          {x-start, x-end, y-start, y-end}:<N | M>:interval>
#      NO   do not show a plot
#      YES  show plot with following specifications
#           BOTH    print raw coordinates into MATHFILE and plot to OUTFILE
#           OUTFILE plot only to OUTFILE
#             LOG   scaling of both axes
#             STD   non-log scaling
#             {...} plot range of both parameters
#             N     use xNm to plot immigration, x=<1,2,3,4>
#                   depending on the inheritance characteristic of the data
#             M     plot migration rate/mutation rate as immigration axis
#             interval the plot range is broken up into interval intervals
plot=NO
#
# Print plot data into a file 
#   Syntax: mathtfile=mathfile the values are printed in a mathematica readable way
mathfile=mathfile
#
#-------------------------------------------------------------------------------
#
# Profile likelihood for each estimated parameter
#   Syntax profile=<NONE | <ALL | TABLES | SUMMARY>:
#               <PRECISE | SPLINE | DISCRETE | QUICKANDDIRTY | FAST>  >
#      NONE    do not calculate profile likelihoods
#      ALL     print individual profile tables and summary [default]
#      TABLES  show only tables and no summary
#      SUMMARY show only summary
#           PRECISE  evaluate profile likelihood at percentiles [Default]
#           SPLINE   use splines to calculate precentile values [DOES NOT WORK!]
#           FAST     assumes that there is no interaction of parameters
#           QUICKANDDIRTY same as FAST except in last calculation cycle assumes interaction
#           DISCRETE .........
profile=ALL:PRECISE
#
#-------------------------------------------------------------------------------
#
# Print tree into treefile
#   Syntax print-tree=< NONE | ALL | BEST | LASTCHAIN >
#         NONE no tree printed [Default, and only choice using parallel
#         ALL  print all visited genealogies [careful this will be huge]
#         BEST print only the best tree visited
#         LASTCHAIN print all trees in last chain
print-tree=NONE
#
#-------------------------------------------------------------------------------
#
# write intermediate minimal statistics into a file for later use
#   Syntax write-summary=<NO | YES:SUMFILE >
#                Default is NO, with YES the user needs to 
#                give a file to record the summary statistics
write-summary=NO
#
#-------------------------------------------------------------------------------
#
# Likelihood ratio test
#   Syntax l-ratio=<NO | YES:values_to_test>
#       Values_to_test are compared to the values generated in the run
#   values_to_test={ab..bbab..ba ... a}
#        the {} is a square matrix with values for the population sizes
#        on the diagonal and migration rates off-diagonal
#        the values a for the diagonal can be any of these:
#        number  constant, the value is for example 0.002
#        *       free to vary, the default is * for every parameter
#        m       mean of theta, this can be a subgroup of all thetas
#                for example the theta 1-3 are averaged and thetas 4,5 are estimated
#        the values b for the migration rates can be any of these:
#        number  constant, the value is for example 45.0 or 0.0
#        *       free to vary, the default is * for every parameter
#        m       mean of M_ij, this can be a subgroup of migration rates
#                for example the M_1-3i are averaged and M_4,5i are estimated
#        M       means of 4Nm (diploid), 2Nm (haploid), Nm (mtDNA, Y-chromosome)
#        s       symmetric migration rates M
#        S       symmetric migrants 4Nm
#        an example for 5 populations could look like this:
#        l-ratio=YES:{*s00s s*s00 0s*s0 00s*s s00s*
#        this describes a circular stepping stone model with 5 symmetric rates
#         and independent sizes, a very basic stepping stone with 2 parameters would
#        look like this l-ratio=YES:{mm00m mmm00 0mmm0 00mmm m00mm}
#        [The L-RATIO statement can be repeated]
#  Default: l-ratio=NO
#
#-------------------------------------------------------------------------------
#
# AIC model selection [do not use yet, will come in Summer 2004]
#   Syntax aic-modeltest=<NO | YES:<FAST | EXHAUSTIVE>>
#       FAST        [do not use yet]
#       EXHAUSTIVE  [do not use yet]
aic-modeltest=NO
#
#-------------------------------------------------------------------------------
#
# Print a histogram of the time of migration events for each M(i->j)
mig-histogram=ALL:mighistfile
skyline=YES:0.100000:skylinefile      #needs mig-histogram
#
#
################################################################################
# Parameter start settings
################################################################################
#
#   Syntax: theta=<FST | OWN:<{value} | {value1, value2, ...., valuen} | RANDOM:{mean std}>
#      migrationt=<FST | OWN:<{value} | {value1, value2, ...., valuen} | RANDOM:{mean std}>
#        FST     starting parameter are derived from
#                an FST-like calculation (Beerli&Felsenstein 1999
#        OWN     starting values are supplied by user
#           {value}   if only one value is supplied then all population
#                     have the same starting value
#           {value1, value2, ..., valuen} each population has its
#                     own starting value, if the number of values is
#                     insuffient, then the last value is the template
#                     for the remaining populations
#        RANDOM  starting parameter is drawn randomely from a Normal distribution
#           {mean std} with mean and standard deviation
theta=FST
migration=FST
#
#-------------------------------------------------------------------------------
# Mutation rate modifiers
#
#   Syntax: mutation=<NOGAMMA | GAMMA:alpha | OWN:loci: rate1 rate2 ... rate_loci>
#      NOGAMMA      all loci have same mutation rate
#      GAMMA:alpha  mutation rate has Gamma distribution with alpha
#      OWN          mutation rate is different for every locus, but fixed
#         :loci: rate1, ...     number of loci, rate of locus 1, locus 2 etc.
#      DATA         mutation rate modifier is deducted from loci in the data
#                   using Watterson's Theta and then scaling all rates Theta_locus/mean(Theta_loci
mutation=DATA#
#-------------------------------------------------------------------------------
# FST model
#
fst-type=THETA
#
#-------------------------------------------------------------------------------
# Custom migration model
#
#    Syntax: custom-migration={ab..bbab..ba ... a}
#        the {} is a square matrix with values for the population sizes
#        on the diagonal and migration rates off-diagonal
#        the values a for the diagonal can be any of these:
#        c       constant, the value needs to be defined in the theta option
#        *       free to vary, the default is * for every parameter
#        m       mean of theta, this can be a subgroup of all thetas
#                for example the theta 1-3 are averaged and thetas 4,5 are estimated
#        the values b for the migration rates can be any of these:
#        c       constant, the value needs to be defined in the theta option
#        *       free to vary, the default is * for every parameter
#        m       mean of M_ij, this can be a subgroup of migration rates
#                for example the M_1-3i are averaged and M_4,5i are estimated
#        M       means of 4Nm (diploid), 2Nm (haploid), Nm (mtDNA, Y-chromosome)
#        s       symmetric migration rates M
#        S       symmetric migrants 4Nm
#        an example for 5 populations could look like this:
#        custom-migration={*s00s s*s00 0s*s0 00s*s s00s*
#        this describes a circular stepping stone model with 5 symmetric rates
#         and independent sizes, a very basic stepping stone with 2 parameters would
#        look like this custom-migration={mm00m mmm00 0mmm0 00mmm m00mm}
custom-migration={****************}
#
# Influence of geography on migration rate
# a distance matrix between populations changes the migration rate matrix so that
# (genetic?) migration rates =  inferred migration rate / distance ~ a dispersion coefficient
# the geofile contains a number of populations, names for populations (10 characters), they
# need to be in order of the dataset. And the distances between the populations, they do not
# need to be symmetric
#    Syntax: geo:<NO | YES:filename>
#             NO       distances among populations are considered to be 1 [all equal]
#             YES      distances are read from a file
geo=NO
#
#
################################################################################
# Search strategies
################################################################################
#
# MCMC Strategy method
#    Syntax: bayes-update=< NO | YES>
#        NO      maximum likelihood method
#        YES     Bayesian method [will come in version 2.0]
# Some of the options are only available in one or other mode
# BAYESIAN OPTIONS
#        bayes-updatefreq=VALUE 
#            VALUE      is a ratio between 0 and 1
#                       ratio of how many times the genealogy is updated compared to the parameters
#                       If the value is 0.4 in a 2 population scenario and with 1000000 steps
#                       The tree will be evaluated 400000 times, Theta_1, Theta_2, M_21, and M_12
#                        will be each evaluated 125000 times.
#        bayes-posteriorbins=VALUE VALUE
#            VALUE      is the number of bins in the psterior distribution histogram for Theta or M
#        bayes-posteriormaxtype=< ALL | P99 | MAXP99 | P100 >
#            ALL        plots the WHOLE prior-parameter range

#            P99        plots from the minimum prior range value to

#                       the 99% percentile value of EACH parameter

#            MAXP99     sets all axes from minimum to the maximal

#                       99% percentile value of ALL parameter

#            P100       plots from the minimum prior range value to

#                       the 100% percentile value of EACH parameter

#        bayes-file=<YES|NO>:FILENAME
#            FILENAME is the name of the file that will contain
#                    the results for the posterior distribution
#        bayes-allfile=<YES|NO>:FILENAME
#            FILENAME is the name of the file that will contain
#                    all parameters of the posterior distribution [HUGE]
#        bayes-allfileinterval=INTERVAL
#            INTERVAL is the interval at which all parameters are written to file

#        
#        bayes-proposals= THETA < SLICE | METROPOLIS >
#        bayes-proposals= MIG < SLICE | METROPOLIS >
#               SLICE uses the slice sampler to propose new parameter values
#               METROPOLIS uses the Metropolis-Hastings sampler
#               (this is done for each parameter group: THETA or MIGration)
#        
#        bayes-priors= THETA <UNIFORM unipriorvalues | EXP exppriorvalues | WINDOWEXP wexppriorvalues | MULT multpriorvalues | FIXED>
#        bayes-priors= MIG <UNIFORM unipriorvalues | EXP exppriorvalues | WINDOWEXP wexppriorvalues | MULT multpriorvalues | FIXED>
#                unipriorvalues: min max
#                exppriorvalues: min mean max
#                wexppriorvalues: min mean max delta
#                multpriorvalues: min max mult
#                fixed = no value is changed during the run
#
# Maximum likelihood OPTIONS
#        short-chains=VALUE   VALUE is 1..n [Default is 10]
#        short-inc=VALUE      VALUE is the number of updates that are not recorded
#        short-sample=VALUE   VALUE is the number of sampled updates
#
# Search OPTIONS for both strategies
#        long-chains=VALUE   VALUE is 1..n [Default is 3]
#        long-inc=VALUE      VALUE is the number of updates that are not recorded
#        long-sample=VALUE   VALUE is the number of sampled updates
#        burn-in=VALUE       VALUE is the number of updates to discard at the beginning
#
bayes-update=YES
bayes-updatefreq=0.500000
bayes-posteriorbins=500 500
bayes-posteriormaxtype=TOTAL
bayes-file=YES:bayesfile-dna
bayes-allfile=YES:bayesallfile-dna
bayes-allfileinterval=1
bayes-proposals= THETA METROPOLIS-HASTINGS Sampler
bayes-proposals= MIG METROPOLIS-HASTINGS Sampler
bayes-priors= THETA UNIFORMPRIOR: 0.000000 0.100000 
bayes-priors= MIG UNIFORMPRIOR: 0.000000 1000.000000 
#
long-chains=1
long-inc=10
long-sample=10000
burn-in=10000
#
#-------------------------------------------------------------------------------
# Schemes to improve MCMC searching
#
# Heating schemes {MCMCMC = MC cubed}
#    Syntax: heating=< NO | <YES | ADAPTIVE>:SKIP:TEMPERATURES
#        NO    No heating
#        YES   heating using TEMPERATURES
#        ADAPTIVE adaptive heating using start TEMPERATURES
#        SKIP skip that many comparisons, this lengthens the run by SKIP
#            TEMPERATURES    { 1.0, temp1, temp2, temp3 .. tempn}
#     Example: heating=YES:1:{1.0, 1.2, 3.0,6.0}
heating=YES:1:{1.0, 1.2, 3.0,6.0}
#
# Lengthening chain schemes
#    Syntax: moving-steps=< NO | YES:VALUE>
#       VALUE   frequency is between 0..1
moving-steps=NO
#
#    Syntax: long-chain-epsilon=VALUE
#       VALUE    is between 0..INFINITY
#                the VALUE is the likelihood ratio between the old and thew chain
#                the VALUE depends on the number of parameters: with 1 values of 0.5 are great
#                but with many parameters values and bad data >20 is more reasonable
long-chain-epsilon=INFINITY
#
#    Convergence statistic [Gelman and Rubin]
#    Syntax: gelman-convergence=< YES:Pairs|Summary | NO >
#       NO      do not use Gelman's convergence criterium
#       YES     use Gelman's convergence criteria between chain i, and i-1
#               PAIRS reports all replicate pairs
#               SUM   reports only mean and maxima
gelman-convergence=Yes:Sum
#
#    Syntax: replicate=< NO | YES:<VALUE | LastChains> >
#       NO     no replication of run
#       YES    replicate run
#           VALUE     number between 2 and many, complete replicates
#           LastChains  replications over last chains
replicate=YES:5
#
# Migration rates are attracted to zero (fatal attraction)
# Resistance is the lowest migration value for all but the last chain
#    Syntax resistance=VALUE
#        VALUE is the lowest migration rate value allowed during all but the last chain
#              typical values are 0.01 or _lower_ for data with sequences and 0.0001 or _lower_ for other data
resistance=0.000100
#
#-------------------------------------------------------------------------------
#
end
