Skip to main content

Table 2 Prediction performance in terms of Pearson’s correlation reported by McDowell (2016); A). Prediction performance in terms of Pearson’s correlation reported by Bellot et al. (2018); B) for traits height and heel bone mineral density. In set “BEST,” the 10 k or 50 k were chosen the top most-associated SNPs, with k = 1000, with the lowest P-values in a GWAS on the TRN set for each trait. In set “UNIF,” the genome was split in windows of equal physical length and the most associated SNP within each window was chosen. MLP denotes multilayer perceptron and CNN convolutional neural networks

From: A review of deep learning applications for genomic selection

A Species Trait OLS RR LR ER BRR MLP
  Arabidopsis Dry Matter 0.36 0.4 0.4 0.42 0.39 0.4
   Flowering 0.8 0.82 0.83 0.82 0.82 0.86
  Maize Flowering 0.22 0.33 0.32 0.33 0.32 0.35
   GY 0.47 0.59 0.49 0.51 0.57 0.55
  Wheat SGN 0.15 0.27 0.33 0.36 0.28 0.33
   TYM 0.59 0.61 0.74 0.73 0.64 0.76
B Species Trait Method 10kBEST 10kUNIF 50kBEST 50kUNIF  
  Human Height BayesB 0.47 0.38 0.48 0.42  
   Height BRR 0.47 0.37 0.47 0.39  
   Height MLP 0.45 0.36 0.45 0.39  
   Height CNN 0.44 0.34 0.42 0.29  
   HBMD BayesB 0.28 0.22 0.26 0.24  
   HBMD BRR 0.28 0.21 0.24 0.22  
   HBMD MLP 0.15 0.11 0.07 0.09  
   HBMD CNN 0.27 0.18 0.10 0.11  
  1. SGN spike grain number; TYM Time young microspore and HBMD Heel bone mineral density