drafting the main function

NAMESPACE

@@ -9,8 +9,14 @@ export(fn_G_numeric_to_non_numeric)
 export(fn_G_split_off_alternative_allele)
 export(fn_G_to_vcf)
 export(fn_classify_allele_frequencies)
+export(fn_cross_validation_across_populations_bulk)
+export(fn_cross_validation_across_populations_lopo)
+export(fn_cross_validation_across_populations_pairwise)
+export(fn_cross_validation_preparation)
+export(fn_cross_validation_within_population)
 export(fn_cv_1)
 export(fn_elastic_net)
+export(fn_estimate_memory_footprint)
 export(fn_filter_genotype)
 export(fn_filter_phenotype)
 export(fn_gBLUP)
@@ -24,4 +30,5 @@ export(fn_ridge)
 export(fn_save_genotype)
 export(fn_save_phenotype)
 export(fn_simulate_data)
+export(fn_subset_merged_genotype_and_phenotype)
 export(fn_vcf_to_G)

R/metrics.R

@@ -8,6 +8,7 @@
 #'
 #' @param y_true numeric vector observed phenotype values
 #' @param y_pred numeric vector predicted phenotype values
+#' @param verbose show genomic prediction performance metric calculation messages? (Default=FALSE)
 #' @returns
 #' Ok:
 #'  $mbe: mean bias error
@@ -69,7 +70,7 @@ fn_prediction_performance_metrics = function(y_true, y_pred, verbose=FALSE) {
     } else {
         r2 = 1 - (sum(error^2, na.rm=TRUE) / sum((y_true-mean(y_true, na.rm=TRUE))^2, na.rm=TRUE))
     }
-    corr = suppressWarnings(stats::cor(y_true, y_pred, method="pearson"))
+    corr = suppressWarnings(stats::cor(y_true, y_pred, method="pearson", use="na.or.complete"))
     ### Power to select true top and bottom 10%
     n_top_or_bottom_10 = max(c(1, round(0.1*n)))
     top10_dec_true = order(y_true, decreasing=TRUE)[1:n_top_or_bottom_10]

R/models.R

@@ -104,6 +104,8 @@ fn_ols = function(list_merged, vec_idx_training, vec_idx_validation, other_param
     }
     X_training = list_merged$G[vec_idx_training, ]
     y_training = list_merged$list_pheno$y[vec_idx_training]
+    X_training = X_training[!is.na(y_training), ] ### Remove missing phenotype data from the training set
+    y_training = y_training[!is.na(y_training)] ### Remove missing phenotype data from the training set
     X_validation = list_merged$G[vec_idx_validation, ]
     y_validation = list_merged$list_pheno$y[vec_idx_validation]
     ### Adding covariate to explanatory matrix
@@ -275,6 +277,8 @@ fn_ridge = function(list_merged, vec_idx_training, vec_idx_validation, other_par
     }
     X_training = list_merged$G[vec_idx_training, ]
     y_training = list_merged$list_pheno$y[vec_idx_training]
+    X_training = X_training[!is.na(y_training), ] ### Remove missing phenotype data from the training set
+    y_training = y_training[!is.na(y_training)] ### Remove missing phenotype data from the training set
     X_validation = list_merged$G[vec_idx_validation, ]
     y_validation = list_merged$list_pheno$y[vec_idx_validation]
     ### Adding covariate to explanatory matrix
@@ -416,6 +420,8 @@ fn_lasso = function(list_merged, vec_idx_training, vec_idx_validation, other_par
     }
     X_training = list_merged$G[vec_idx_training, ]
     y_training = list_merged$list_pheno$y[vec_idx_training]
+    X_training = X_training[!is.na(y_training), ] ### Remove missing phenotype data from the training set
+    y_training = y_training[!is.na(y_training)] ### Remove missing phenotype data from the training set
     X_validation = list_merged$G[vec_idx_validation, ]
     y_validation = list_merged$list_pheno$y[vec_idx_validation]
     ### Adding covariate to explanatory matrix
@@ -557,6 +563,8 @@ fn_elastic_net = function(list_merged, vec_idx_training, vec_idx_validation, oth
     }
     X_training = list_merged$G[vec_idx_training, ]
     y_training = list_merged$list_pheno$y[vec_idx_training]
+    X_training = X_training[!is.na(y_training), ] ### Remove missing phenotype data from the training set
+    y_training = y_training[!is.na(y_training)] ### Remove missing phenotype data from the training set
     X_validation = list_merged$G[vec_idx_validation, ]
     y_validation = list_merged$list_pheno$y[vec_idx_validation]
     ### Adding covariate to explanatory matrix
@@ -747,8 +755,8 @@ fn_Bayes_A = function(list_merged, vec_idx_training, vec_idx_validation,
 }
 
 #' Bayes B regression model 
-#'  (scaled t-distributed effects with probability $\pi$; and zero effects with probability $1-\pi$, 
-#'  where $\pi \sim \beta(\theta_1, \theta_2)$)
+#'  (scaled t-distributed effects with probability \eqn{\pi}; and zero effects with probability \eqn{1-\pi}, 
+#'  where \eqn{\pi \sim \beta(\theta_1, \theta_2)})
 #'
 #' @param list_merged list of merged genotype matrix, and phenotype vector, as well as an optional covariate matrix
 #'  $G: numeric n samples x p loci-alleles matrix of allele frequencies with non-null row and column names.
@@ -889,8 +897,8 @@ fn_Bayes_B = function(list_merged, vec_idx_training, vec_idx_validation,
 }
 
 #' Bayes C regression model
-#'  (normally distributed effects ($N(0, \sigma^2_{\beta})$) with probability $\pi$; and zero effects 
-#'  with probability $1-\pi$, where $\pi \sim \beta(\theta_1, \theta_2)$)
+#'  (normally distributed effects (\eqn{N(0, \sigma^2_{\beta})}) with probability \eqn{\pi}; and zero effects 
+#'  with probability \eqn{1-\pi}, where \eqn{\pi \sim \beta(\theta_1, \theta_2)})
 #'
 #' @param list_merged list of merged genotype matrix, and phenotype vector, as well as an optional covariate matrix
 #'  $G: numeric n samples x p loci-alleles matrix of allele frequencies with non-null row and column names.