Comparision of GS Models

From HORTS 1993
Jump to navigation Jump to search

The scope

  • Various GS models such as Bayes B, SPLS, BL, RR_BLUP, PLS have been compared[1]
  • Prediction accuracy in Time to young microspore: 0.5 to 0.84
  • Prediction accuracy of grain number under Osmatic stress conditions : 0.27 - 0.46
  • Prediction accuracy of grain number under controlled conditions  : 0.1 - 0.42
  • Accuracies were generally lower in independent validation than cross validation (?!)
  • BayesB and SPLS were turned out to be the best models
  • Hence BayesB and SPLS capture the LD between the markers and traits => higher accuracy.
  • Excluding markers from QTL mapping reduces the accuracies (?!)

==