ADMIX
2.0
Computing admixture coefficients from molecular
data
The program Admix 2.0 can be used to compute the
estimators mY initially described in Bertorelle & Excoffier (1998) and
extended to any number of parental populations by Dupanloup & Bertorelle
(2001).
Admix 2.0 runs on PC. The Apple version is not available
yet. The number of different alleles for each locus is restricted to 1000 !
- 1st line: time in years to
the admixture event (If this time is not known, one must write 0 instead !)
- 2nd line: mutation rate per
year (If a time to the admixture event is given in the preceding line, one
should indicate a mutation rate !)
- 3rd line: number of
parental populations
- 4th line: number of loci
- Following lines: for each
loci:
- 1st line: name of the
locus (this line must contain at least one character !)
- 2nd line: number of
different alleles (k) observed at that locus
- Next k lines: lower
triangular matrix specifying the molecular distances between alleles. In the
case of DNA sequences or RFLPs, the molecular distance between 2 alleles is
just the number of substitutions. In case of microsatellite data, the
molecular distance is the squared difference in allele size (see Bertorelle
& Excoffier for more details).
- Following lines: for each
loci:
- 1st line: name of the
locus (this line must contain at least one character !)
- 2nd line: sample sizes
(number of chromosomes) in the admixed and the parental populations at that
locus. The first number refers to the admixed population. The following
numbers refer to the parental populations.
- Next k lines: for each
locus, for each population, the number of copies observed of each allele.
Again, the first number refers to the admixed population, and the following
numbers refer to the parental populations.
A possible data file for 2 parental populations and 3 loci
could therefore look like:
1000
0.00005
2
3
LocA
4
0
1 0
4 1 0
9 4 1 0
LocB
5
0
1 0
4 1 0
9 4 1 0
25 9 4 1 0
LocC
2
0
1 0
LocA
30 24 27
3 6 0
3 12 3
15 3 15
9 3 9
LocB
51 30 36
9 0 12
9 6 0
30 6 21
0 9 0
3 9 3
LocC
30 30 24
12 21 6
18 9 18
- result_admix.txt This output file will contain
the estimated admixture coefficients. The admixture coefficients correspond to
the relative contribution to the hybrid population from the parental
populations given in the same order as were declared the populations in the
input file. If you requested it, this output file will also provide the
bootstrap averages and bootstrap standard deviations of the admixture
coefficients. These last couples of values are computed by randomly resampling
from the original data sets.
- bootstrap.txt If you requested the estimation
of bootstrap averages and bootstrap standard deviations of the admixture
coefficients, this file will contain the N sets of boostrapped values over N
bootstraps operations.
- log_file_admix.txt This file gives all the
details on the computations done by Admix 2.0. It contains all the run-time
WARNINGS and ERRORS encountered during these computations.
In case of any trouble using Admix 2.0, please contact I.
Dupanloup, Institute of Ecology and Evolution, University of Bern. E-mail: isabelle.dupanloup@zoo.unibe.ch
August 2002: When 2 parental populations have the same
allelic composition (in the observed samples or in bootstrapped samples), the
software crashed. Now the program stops and a message error is given in the log
file. Note that this problem is not linked to our software. It is impossible to
estimate admixture proportions when 2 (or more) parental populations are
genetically identical. Practically, one shall esclude 1 (or more) samples or
pool the samples.
- Bertorelle, G. and Excoffier, L. (1998) Inferring admixture proportions
from molecular data. Molecular Biology and Evolution 15(10): 1298-1311.
- Dupanloup, I. and Bertorelle, G. (2001) Inferring admixture proportions
from molecular data: extension to any number of parental populations.
Molecular Biology and Evolution 18(4): 672-675.
Isabelle
Dupanloup, Institute of Ecology and Evolution, University of Bern