[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

*To*: "Hanson W. Wei" <weiw@whio.lincoln.ac.nz>*Subject*: Re: singularities*From*: bsouthey@iastate.edu*Date*: Tue, 07 Dec 1999 17:58:36 CST*cc*: Arthur Gilmour <gilmoua@ornsun.agric.nsw.gov.au>, asreml@chiswick.anprod.csiro.au*In-reply-to*: Your message of "Tue, 07 Dec 1999 12:46:25 +1300." <0FMC00NS3FGPV9@kauri.lincoln.ac.nz>*Sender*: asreml-owner@lamb.chiswick.anprod.csiro.au

Hi, How many animals do you have? Either you did not send all 150 or you told ASREML incorrectly. Actually you only have 106 unique tags with 125 observations. Thus, the most of your other singularities are coming because most animals only have one record (except the COR's that have 2 excluding COR1). For a single trait fixed effects model of the mean, breed and age you should get 2 singularities due to the dependencies i.e. X is not full rank. In a more complex model with interactions you will get more because of two completely missing cells. One should also be careful because you only have 5 observations for breed 2. For a bivariate analysis, you should get 4 singularities. If you leave the number of animals at 150 then ASREML reports 48. Where did the extra 44 come from? Try 150 minus 106 or the animals that don't have any observations in the data. With essentially one observation per animal you will need to provide an A matrix or equivalent. Bruce -- Asreml mailinglist archive: http://www.chiswick.anprod.csiro.au/lists/asreml

**References**:**singularities***From:*"Hanson W. Wei" <weiw@whio.lincoln.ac.nz>

- Prev by Date:
**Re: Standard error of heritability.** - Next by Date:
**Christmas greetings** - Prev by thread:
**singularities** - Next by thread:
**Christmas greetings** - Index(es):