From jillian.maddoxalumni.unimelb.edu.au  Mon Jan  7 19:53:32 2013
From: Jill Maddox <jillian.maddoxalumni.unimelb.edu.au>
To: Multiple Recipients of <crimap-usersanimalgenome.org>
Subject: Family sizes - split into sub-families for faster CRI-MAP
Date: Mon, 07 Jan 2013 19:53:32 -0600

Hi All

The memory requirements for CRI-MAP are less for smaller family sizes.
Depending on the family structure it can be advantageous to split a family
into sub-families providing the splitting does not duplicate phased
chromosomes. This is particularly useful with large full-sibling families
as the memory requirements are much greater for these.

Examples of family structures that are good for splitting CEPH-like - 3-gen
full-sibling family where both parents are single children (if either
parent has genotyped siblings then splitting will duplicate some phased
chromosomes). Split into smaller 3-gen full-sibling families (each new
family should have genotypes for grandparents and parents as well as
progeny). Note, the number of terminal progeny per family needs to be large
enough for genotype frequencies to be approx HWE - maybe try 30+ terminal
progeny per family.

Large 2-gen families - half-sibling or full-sibling (2-generation means no
duplication of phased chromosomes so easy to split). Maybe try 100+
terminal progeny per half-sib family and 30+ per full-sib family.

Any pedigree where parents have siblings or there are more than 3-
generations or complex interrelationships needs to be split more carefully
so as to minimise the number of duplicated phased chromosomes or genotype
wastage.

I am developing a module for CRI-MAP to make splitting easier. Please note
that the Monsanto CRIGEN program does not split families correctly for many
family types.

Regards

 
Jill 

 
*************************************************************** 

Jill Maddox 16 Park Square Port Melbourne, 3207 Australia phone: 03 9646
0428 E-mail: jillian.maddoxalumni.unimelb.edu.au

***************************************************************