|
Activity Number:
|
426
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 9, 2006 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #306073 |
|
Title:
|
Estimating population structure
|
|
Author(s):
|
Suvajit Samanta*+ and Bruce S. Weir
|
|
Companies:
|
North Carolina State University and University of Washington
|
|
Address:
|
3223 Octavia Street, Raleigh, NC, 27606,
|
|
Keywords:
|
population structure ; f-statistics ; moment estimator ; sampling error
|
|
Abstract:
|
The probabilities that two, three, four or two-pairs of alleles taken randomly from a population are identical by descent are parameters of great importance for population genetic theory. Only for the two-allele parameter is there a well-established estimation procedure. Moment estimation has been used when the distribution of allele frequencies over populations is not known, and maximum likelihood estimates have been given under the assumptions of normal or Dirichlet distributions. We now propose a moment estimator for the three-allele probability and we show that the large variances can be reduced by combining information over many loci. By comparing estimates of two- and three-allele probabilities we can address the relevance of the normal or Dirichlet distributions since the relationship between the two parameters is known for those distributions.
|
