2.4 GS

It’s required to provide genotype and phenotype files to run GS in iPat. User-provided covariates will serve as fixed effects in the GS model, but none of GS methods will take user-defined kinship in iPat.

In iPat, users can perform two types of GS. If the ‘validation’ option is unchecked, iPat would estimate genetic estimated breeding values (GEBV) for samples in the whole dataset based on the phenotyped samples. Whereas when users check the option ‘Validation on accuracy’, iPat will split whole dataset into training and testing samples in order to verify the capability of trained GS model. For this manner, validation accuracies will be reported instead of GEBVs.

## Column 1 ['V1'] of item 2 is missing in item 1. Use fill=TRUE to fill with NA (NULL for list columns), or use.names=FALSE to ignore column names. use.names='check' (default from v1.12.2) emits this message and proceeds as if use.names=FALSE for  backwards compatibility. See news item 5 in v1.12.2 for options to control this message.
## Column 1 ['V1'] of item 2 is missing in item 1. Use fill=TRUE to fill with NA (NULL for list columns), or use.names=FALSE to ignore column names. use.names='check' (default from v1.12.2) emits this message and proceeds as if use.names=FALSE for  backwards compatibility. See news item 5 in v1.12.2 for options to control this message.
## Column 1 ['V1'] of item 2 is missing in item 1. Use fill=TRUE to fill with NA (NULL for list columns), or use.names=FALSE to ignore column names. use.names='check' (default from v1.12.2) emits this message and proceeds as if use.names=FALSE for  backwards compatibility. See news item 5 in v1.12.2 for options to control this message.

TABLE 2.4: Arguments for GS
Arguments	Default	Options	Description
Validation on accuracy	unchecked	checked, unchecked	If k-fold validation is performed in GS
Folds	3	3, 5, 10	K value for K-fold validation. For example, if k = 5, 20% samples will be tested as accuracy validation.
Iteration	10	Range from 10 to 100	Number of iterations. Each iteration will split whole dataset into training part and testing part based on the specified k value for K-fold validation.

Output files from GS include:

Figures: GEBV distribution, Validation accuracies, Predicted versus observed values, Heterozygosity distribution, and Phenotype overview.
Tables: GEBV report, Marker effects, Validation results, and Predicted values

2.4.1 gBLUP and rrBLUP

There’s no specific arguments available for both of these methods.

Reference

Endelman,J. (2011) Ridge regression and other kernels for genomic selection in the R package rrBLUP. Plant Genome, 4, 250–255.

2.4.2 BGLR

## Column 1 ['V1'] of item 2 is missing in item 1. Use fill=TRUE to fill with NA (NULL for list columns), or use.names=FALSE to ignore column names. use.names='check' (default from v1.12.2) emits this message and proceeds as if use.names=FALSE for  backwards compatibility. See news item 5 in v1.12.2 for options to control this message.
## Column 1 ['V1'] of item 2 is missing in item 1. Use fill=TRUE to fill with NA (NULL for list columns), or use.names=FALSE to ignore column names. use.names='check' (default from v1.12.2) emits this message and proceeds as if use.names=FALSE for  backwards compatibility. See news item 5 in v1.12.2 for options to control this message.
## Column 1 ['V1'] of item 2 is missing in item 1. Use fill=TRUE to fill with NA (NULL for list columns), or use.names=FALSE to ignore column names. use.names='check' (default from v1.12.2) emits this message and proceeds as if use.names=FALSE for  backwards compatibility. See news item 5 in v1.12.2 for options to control this message.

TABLE 2.5: Arguments for BGLR
Arguments	Default	Options	Description
Model for predictors	BRR	BRR, Bayes A, Bayes B, Bayes C, BL	Model to estimate marker effects
nIter	5k	Range from 1k to 100k	Number of iteration
burnIn	500	Range from 200 to 10k	Burn-in values

Reference

Pérez, Paulino. et al. (2014) Genome-wide regression and prediction with the BGLR statistical package. Genetics, 198, 483-495.