Microsats question

[Julio]

Hi
I'm running a microsats datasets with the following parameters:
but std (r) is consistently 2-3x the mean(r). I'm not using priors,
and running K2-4. I don't think it it converging.
Any help would be appreciated,
Regards,
Julio



#define NUMLOCI 37 // (int) number of loci in data file
#define LABEL 1 // (B) Input file contains individual labels
#define POPDATA 0 // (B) Input file contains a population
identifier
#define POPFLAG 0 // (B) Input file contains a flag which says
whether to use popinfo when
USEPOPINFO==1
#define PHENOTYPE 1 // (B) Input file contains phenotype
information
#define EXTRACOLS 0 // (int) Number of additional columns of data
before the genotype data start.
#define PHASEINFO 0 // (B) the data for each individual contains a
line
indicating phase
#define MARKOVPHASE 0 // (B) the phase info follows a Markov model.

#define MISSING 0 // (int) value given to missing genotype data
#define PLOIDY 2 // (int) ploidy of data

#define ONEROWPERIND 1 // (B) store data for individuals in a single
line
#define MARKERNAMES 1 // (B) data file contains row of marker names
#define MAPDISTANCES 1 // (B) data file contains row of map
distances
// between loci

Program Parameters

#define MAXPOPS 2 // (int) number of populations assumed
#define BURNIN 10000 // (int) length of burnin period
#define NUMREPS 50000 // (int) number of MCMC reps after burnin
#define FREQSCORR 0 // (B) allele frequencies are correlated among
pops
#define ONEFST 0 // (B) assume same value of Fst for all
subpopulations.

#define INFERALPHA 1 // (B) Infer ALPHA (the admixture parameter)
#define POPALPHAS 1 // (B) Individual alpha for each population

#define INFERLAMBDA 0 // (B) Infer LAMBDA (the allele frequencies
parameter)
#define POPSPECIFICLAMBDA 0 //(B) infer a separate lambda for each
pop
(only if INFERLAMBDA=1).

#define NOADMIX 0 (B) Use no admixture model
#define LINKAGE 1 // (B) Use the linkage model model
#define PHASED 0 // (B) Data are in correct phase (required
unless data
are diploid)
If (LINKAGE=1, PHASED=0), then PHASEINFO can be
used--this is an
extra line in the input file that gives phase
probabilities.
When PHASEINFO =0 each value is set to 0.5,
implying
no phase information.

#define LOG10RMIN -4.0 //(d) Log10 of minimum allowed value of r
under
linkage model
#define LOG10RMAX 1.0 //(d) Log10 of maximum allowed value of r
#define LOG10RPROPSD 0.1 //(d) standard deviation of log r in
update
#define LOG10RSTART -2.0 //(d) initial value of log10 r

#define COMPUTEPROB 1 // (B) Estimate the probability of the Data
under
the model. This is used when choosing
the
best number of subpopulations.

#define ADMBURNIN 2500 //(int) initial period of burnin with
admixture