Calculates likelihoods of estimators based on the number of haplotypes and/or
the minimum number of inferred recombination events.  (See Wall 2000 MBE for
details).  To compile, type

cc -O -o hrmpg2 hrmpg2.c mg.c stree.c subs1.c ran1.c -lm

The files my.h and seedms should also be present.

Usage is as follows:

hrmpg2 npop samplesize[1] . . . samplesize[npop] nsites segsites Rmin Rmax pts alpha t0(if alpha!=0.0) (migrate if npop>1) howmany

where these arguments are:
  npop: Number of populations
  samplesize[i]: Number of individuals sampled from the i-th island
  nsites: Number of sites between which recombination can occur (# of bp's) 
  segsites: Number of polymorphic sites in the total sample
  Rmin: Minimum value of 4Nr considered.
  Rmax: Maximum value of 4Nr considered.
  pts: Number of different values of 4Nr considered.
  alpha: Growth rate
  t0: Time (in units of 4N generations, where N is the current pop. size)
      since the start of growth
  howmany: Howmany samples to generate.

Type program name to get a list of command line arguments necessary.  Note
that rec. rates considered are an arithmetic series between the given
minimum and maximum.  Data is generated under an island model of 
subdivision with recent exponential growth.  To generate simple
simulations (no growth and no structure), type, e.g.,

hrmpg2 1 20 1000 10 0. 10. 11 0. 100000

to run simulations for n = 20, 1 kb of sequence, S = 10, and 100,000 trials
each of C = 0.0, 1.0, 2.0, . . . 10.0.

Output consists of the command line arguments followed by the following 
variables and arrays:

  likW[a] = value of 4Nr that maximizes the likelihood of observing exactly 
            a different haplotypes.

  likRM[a] = value of 4Nr that maximizes the likelihood of observing RM = a.

  likRMW[a][b] = value of 4Nr that maximizes the likelihood of observing
                 RM = a and W = b.

  hap[a][b] = number of trials with 4Nr = Rmin + a * (Rmax - Rmin) / (pts - 1)
              with exactly b different haplotypes.

  RM[a][b] = number of trials with 4Nr = Rmin + a * (Rmax - Rmin) / (pts - 1)
             with RM = b.

  RMhap[a][b][c] = # of trials with 4Nr = Rmin + a * (Rmax - Rmin) / (pts - 1)
                   with RM = b and W = c.

If you have questions, contact jdwall@midway.uchicago.edu

Last modified 5/15/00