Differences

This shows you the differences between two versions of the page.

--- covariance_structure_of_correlated_random_effects_for_multiple_traits [2018/03/29 18:40] – created andres
+++ covariance_structure_of_correlated_random_effects_for_multiple_traits [2018/03/29 18:52] – andres
@@ Line 1: / Line 1: @@
 The documentation is  bit confusing on how to declare several correlated effects for several traits. Whereas ''renumf90'' does this automatically for maternal effects models, it is not so obvious if one does it by itself.
 For instance terminal crosses models (e.g. [[https://gsejournal.biomedcentral.com/articles/10.1186/s12711-016-0211-3||here]]) use:
-  * trait_pure: animal
+  * trait_pure: **y = u + e** (animal)
-  * trait_crossbred : sire(crossbred)
+  * trait_crossbred : **y = s + e** (sire of crossbred)
-Imagine that we have two such traits. Data file has "animal" in column 4 and "sire" in column 5.
+Thus we have two traits, purebred and crossbred performance, but effect **u** only affects purebreds and effect **s** only affects crossbreds. Data file has "animal" in column 4 and "sire" in column 5.
+<file>
 ...
 0  3084 cross
@@ Line 20: / Line 20: @@
 ../SNPgr/ped.txt
 (CO)VARIANCES
-		Effect 4	Effect 5
+ 0 0 5
-		Trait 1	Trait2	Trait1	Trait2
+0 0 0
-Effect 4	Trait 1	Var(u)	0 	0	Cov(u,s)
+0 0 12
-	Trait2	0	0	0	0
+</file>
-Effect 5	Trait 1	0	0	0	0
-	Trait2	Cov(s,u)			Var(s)
+Where we have a 4x4 matrix for covariances. In fact this corresponds to
+^      ^	^Effect 4^	^Effect 5^      |
+^      ^	^Trait 1 ^Trait2^Trait1	 ^Trait2|
+^Effect 4^Trait 1|Var(u)|     0 |	0|Cov(u,s)|
+^	^Trait 2 |0	|     0 |	0|	0|
+^Effect 5^Trait 1|0	|     0 |      0|	0|
+^	^Trait 2|Cov(s,u)|	0|	0|Var(s)|