29 populations, 9 loci, 87 alleles... Takes ages?

[ramon...@hotmail.com]

Hi everyone!

I'm new in the group and, even when I made some successful analysis with mtDNA, the first time I wanted to run it for micros is being no so good as I expected.

It seems the infile is right, but the problem is maybe with the settings, since it's taken ages (13 days when we interrupted it) in a parallel analysis with 10 processors.

Here are the settings... What can be wrong?

Thanks a lot in advance.

Ramón.

################################################################################

# Parmfile for Migrate 3.6.4-2160x [do not remove these first TWO lines]

# generated automatically on

# 06/11/15 11:49:25

#

# please report problems to Peter Beerli

#  email: bee...@fsu.edu

#  http://popgen.sc.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

#                  [singlestep versus multistep model, see micro-submodel option]

# 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-submodel=<1|2:{tune,pinc}>

#                 1 means singlestep mutation model (this is the default and the standard

#                 2 is the Multistep model (see Watkins 2007 TPB, section 4.2) it needs

#                   two parameters: tune specifies how close the model is to a singlestep model

#                   so tune=0 --> singlestep, tune=1 --> infinite allele model;

#                   the second parameter defines the probability that the repeat number

#                   is increasing, this value cannot be larger than 0.666, I suggest 0.5.

#                   Example: micro-submodel=2:{0.5,0.5}

#          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

#          inheritance-scalars={values for each locus}

#                these values are multiplied with Theta, for example having

#                two autosomal and a locus on X- and one on Y-chromosome we would give 

#                inheritance-scalars={1 1 0.75 0.25}

#                [if all loci have the same scalar, just use {1}, even for many loci]]

#          population-relabel={assignment for each location in the infile}

#                example is population-relabel={1 2 2}

#          random-subset=number<:seed>

#                allows to subset the dataset randomly, if number > sample in population

#                all samples are taken, if number is smaller then the pop sample is shuffled and

#                and the first number samples are taken.

#                the random number seed guarantees that the

#                same subset is chosen in different runs

#          usertree=<NO | UPGMA | AUTOMATIC | 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=BrownianMicrosatelliteData

include-unknown=NO

inheritance-scalars={1.00000000000000000000}

population-relabel={1}

random-subset=7:64564

usertree=RANDOMTREE

#

################################################################################

# Input options

################################################################################

#

# input file location

#   Syntax infile=FILEPATH

infile=infile

#

# 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=AUTO #OWN:358503883

#

# Specify the title of the run, will be overridden by title in datafile

#    Syntax: title=title text [up to 80 characters]

title=AUTO 

#

#

################################################################################

# 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=NO

#

#-------------------------------------------------------------------------------

#

# 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

#

#-------------------------------------------------------------------------------

#

# Print output to a PDF file [default is outfile.pdf]

#   Syntax pdf-outfile=outfilename.pdf

pdf-outfile=outfile.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 | DISCRETE | QUICK | 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]

#           QUICK    assumes that there is no interaction of parameters

#           FAST     same as QUICK except in last calculation cycle assumes interaction

#           DISCRETE uses fixed mutipliers: 0.02,0.1,0.2,0.5,1,2,5,10,50

profile=ALL:PRECISE

#

#-------------------------------------------------------------------------------

#

# Print tree into treefile

#   Syntax print-tree=< NONE | <ALL | BEST | LASTCHAIN:Increment>:treefile >

#         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

#         with increment INCREMENT

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)

#    Syntax  mig-histogram=<NO | <ALL | MIGRATIONEVENTSONLY>:binsize:mighistfile >

#         NO            do not record any events

#         ALL           record migration and coalescence event

#         MIGRATIONEVENTSONLY record only migration events

#         binsize has to be in mutation units, with an average Theta=0.01 try 0.001

# Print a histogram of the parameters through time (skyline plot)

#    Syntax  skyline=<NO | YES>:binsize:skylinefile >

#         NO            do not calculate parameter estimates through time

#         YES           calculate parameters through time

#         binsize has to be in mutation units, with an average Theta=0.01 try 0.001

#         If the interval is too fine the output will be very noisy

mig-histogram=NO

skyline=NO #needs mig-histogram=ALL:...

#

#

################################################################################

# Parameter start settings

################################################################################

#

#   Syntax: theta=<FST | OWN:<{value} | {value1, value2, ...., valuen} | NRANDOM:{mean std} | URANDOM{min,max}>

#      migrationt=<FST | OWN:<{value} | {value1, value2, ...., valuen} | NRANDOM:{mean std} | URANDOM{min,max}>

#        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

#        NRANDOM  starting parameter is drawn randomely from a Normal distribution

#           {mean std} with mean and standard deviation

#        URANDOM  starting parameter is drawn randomely from a Uniform distribution

#           {min max} with minimum and maximum values

theta=FST

migration=FST

#

#-------------------------------------------------------------------------------

# Mutation rate modifiers

#

#   Syntax: mutation=<NOGAMMA | CONSTANT | ESTIMATE | GAMMA:alpha | OWN:loci: rate1 rate2 ... rate_loci>

#      NOGAMMA      all loci have same mutation rate

#      CONSTANT     all loci have same mutation rate

#      ESTIMATE     BAYESIAN estimate: mutation rate is drawn from prior

#      GAMMA:alpha  ML estimate: 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=CONSTANT

#

# Treatment of inviariant sequence loci

# Syntax: analyze-loci=<A | F | V>

#         A = analyze all loci (Default!)

#         F = analyze all variable loci and ONE invariant and extrapolate

#         V = analyze only variable loci

#analyze-loci=A

#

#-------------------------------------------------------------------------------

# 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 migration 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

# 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:FILENAME|NO>

#            FILENAME is the name of the file that will contain

#                    the results for the posterior distribution

#        bayes-allfile=<<YES|TEMP>:INTERVAL:FILENAME|NO>

#            FILENAME is the name of the file that will contain

#                    all parameters of the posterior distribution [HUGE]

#            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 

#        bayes-priors= MIG <UNIFORM unipriorvalues | EXP exppriorvalues | WINDOWEXP wexppriorvalues 

#                unipriorvalues: min max delta

#                exppriorvalues: min mean max

#                wexppriorvalues: min mean max delta

#

# 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

#        auto-tune=YES:VALUE  VALUE the the target acceptance ratio

#                             if value is missing, it is set to 0.44

#

bayes-update=YES

bayes-updatefreq=0.500000

bayes-posteriorbins=1000 1000

bayes-posteriormaxtype=TOTAL

bayes-file=NO

bayes-allfile=NO

bayes-proposals= THETA METROPOLIS-HASTINGS Sampler

bayes-proposals= MIG METROPOLIS-HASTINGS Sampler

bayes-priors= THETA UNIFORMPRIOR: 0.000000 0.100000 0.010000 

bayes-priors= MIG UNIFORMPRIOR: 0.000000 1000.000000 100.000000 

#

long-chains=1

long-inc=100

long-sample=5000

burn-in=10000  

auto-tune=YES:0.440000

#

#-------------------------------------------------------------------------------

# Schemes to improve MCMC searching and/or thermodynamic integration

#

# 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 [fails sometimes!!]

#        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, 3.0,1000000.0}

# Heating:  swapping chains

#     Syntax: heated-swap=< YES | NO >

#         YES  swapping of chains enabled [DEFAULT]

#         NO   swapping of chains disabled

#      Example: heated-swap=YES

heating=NO

#

# 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=No

#

#    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=NO

#

# 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=0.000001

#

#-------------------------------------------------------------------------------

#

end

[Peter Beerli]

Ramon,

your data seems to be msats (you specify the Brownian model),

you will need to change the priors from

On Jun 21, 2015, at 1:50 PM, ramon...@hotmail.com wrote:

bayes-priors= THETA UNIFORMPRIOR: 0.000000 0.100000 0.010000 

bayes-priors= MIG UNIFORMPRIOR: 0.000000 1000.000000 100.000000 

to

bayes-priors= THETA UNIFORMPRIOR: 0.000000 100.000 10.000 

bayes-priors= MIG UNIFORMPRIOR: 0.000000 100.000000 10.000000 

Aslo your custom-migration suggests that you have 4 populations not not 29 as your subject line, what is it? if it is 29 then it will take forever because of the number of populations (and the number of individuals, how many total?), if it is only 4 populations than it should finish certainly within a day (again depending on the number of loci and individuals).

Peter

[ramon...@hotmail.com]

Peter, 

Thank you very much for your answer.

Yes, I obtained a best-value of 29 populations with sequential SAMOVA, for a total of 525 individuals and 9 loci (for a total of 87alleles). But this method also allows to use significant population structure by giving the software a smaller K value. For these parameters, how many populations do you think would be reasonable?

And regarding the custom-migration, I'm not sure what you talk about...

Thank you, 

Ramón.

[Peter Beerli]

I suggest to have a look at the manual for the custom-migration matrix.

migrate estimates parameters and also allows you to compare models,

in your case the default will be to estimate 841 parameters, this will certainly take a while :-(

but I actually doubt that you can estimate that many parameters given your data, having 841 parameters but only 87 alleles, you may have an interest in looking at the tutorial about model selection (bayes factors) in the tutorial section of the migrate website. I would probably start with 2 or 4 population

Peter

--
You received this message because you are subscribed to the Google Groups "migrate-support" group.
To unsubscribe from this group and stop receiving emails from it, send an email to migrate-suppo...@googlegroups.com.
To post to this group, send email to migrate...@googlegroups.com.
Visit this group at http://groups.google.com/group/migrate-support.
For more options, visit https://groups.google.com/d/optout.

[Monica]

Hi,

I am going to run the analyses for the first time. I am confused about the micro-threshold. The number of repeats in my 10 msats ranges from 20 to 63. In all of them, alleles differ by less than 10 steps. I have 35 populations (20 of them have 20 individuals, 10 have about 10 individuals, and 5 have less than 5 individuals). Should I use the default setting?

Thanks!

[Peter Beerli]

Monica,

I suggest for your first run to use 2 (!) population (pool your populations into two groups) and run the brownian approximation to the microsatellite model, you also need to adjust the prior distribution boundaries.

Peter

[Peter Beerli]

Monica,

set the datatype to Brownian motion, this is the least complicated, fastest way to run microsatellites (the other modes are very slow and not better than Brownian motion)

Peter

On Nov 10, 2015, at 2:14 AM, Monica <monic...@gmail.com> wrote: