User Tools

Site Tools


readme.pedictf90

Differences

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

Link to this comparison view

Next revision
Previous revision
readme.pedictf90 [2018/07/12 21:50]
andres created
readme.pedictf90 [2020/03/13 19:10] (current)
yutaka
Line 1: Line 1:
-**predictf90**+===== predictf90 ​===== 
 + 
 +==== Original description ====
  
 This is program is used to do cross-validations. It reads a blupf90 parameter file, a solutions file This is program is used to do cross-validations. It reads a blupf90 parameter file, a solutions file
-and a data file+and a data file.
 It needs ''​OPTION include_effects''​ followed by a series of effects. It needs ''​OPTION include_effects''​ followed by a series of effects.
 It computes It computes
-y_hat = sum of estimates of the included effects + 
-y_star = y corrected by the other (not included) effects +  * y_hat = sum of estimates of the included effects 
-residual = y - included effects (not a true residual)+  ​* ​y_star = y corrected by the other (not included) effects 
 +  ​* ​residual = y - included effects (not a true residual) 
 + 
 +==== Example ====
  
 For instance consider For instance consider
  
-y = herd + age + animal + e+  ​y = herd + age + animal + e
  
-with ''​OPTION include_effects 3''​+with ''​OPTION include_effects 3''​.
  
-y_hat = animal_hat +  * y_hat = animal_hat 
-ystar = y - herd_hat - age_hat +  ​* ​ystar = y - herd_hat - age_hat
-Which makes cor(y_hat,​y_star) = cor(ebv, corrected y) which is a measure of accuracy+
  
-it outputs the correlation between y_hat and y_star, for instance ​cor(ystar,yhat)=cor(u+euhat)+Which makes cor(y_hat,y_star) = cor(ebvcorrected ywhich is a measure of accuracy.
  
-and outputs these columns into a file, together with animal id (if there is animal in the model) or record number (if not)+It outputs the correlation between y_hat and y_star, for instance cor(ystar,​yhat)=cor(u+e,​ uhat) and outputs these columns into a file, together with animal id (if there is animal in the model) or record number (if not).
  
 In addition, if animal effect is in the model, it produces a file with ebvs from the solutions file. In addition, if animal effect is in the model, it produces a file with ebvs from the solutions file.
 +
 +==== More explanation ====
 +
 +This program splits the phenotype (y) into 3 pieces.
 +
 +  y = (effects to be corrected) + (effects to be kept) + residual
 +
 +In a cross validation, you need a quantity "EBV + residual",​ where EBV is ''​(effects to be kept)''​ shown above, and it is ''​y_star''​.
 +The program calculates the following quantities.
 +
 +  * y_hat: ''​(effects to kept)''​
 +  * y_star: ''​y - (effects to be corrected)''​
 +  * residual: ''​residual''​
 +
 +The user should specify the effects to be kept using ''​OPTION include_effects''​ with the effect numbers.
 +For example, if the breeding value is in the 3rd row in the EFFECTS section in the parameter file, you have to use ''​OPTION include_effects 3''​.
 +
 +==== Output files ====
 +
 +=== yhat_residual ===
 +
 +The main file is ''​yhat_residual'',​ which has corrected phenotypes and predicted residuals.
 +The number of columns in this file depend on the number of traits (N).
 +
 +  * Column 1: Animal ID (renumbered i.e., same as the 1st column in renaddxx.ped)
 +  * Column 2 to N+1: "​y_star"​ explained above
 +  * Column N+2 to 2N+1: "​y_hat"​ explained above
 +  * Column 2N+2 to 3N+1: "​residual"​ explained above
 + 
 +
 +
readme.pedictf90.1531432257.txt.gz · Last modified: 2018/07/12 21:50 by andres