Association of spectroscopically determined leaf nutrition related traits and breeding selection in Sassafras tzumu

Abstract Background Plant traits related to nutrition have an influential role in tree growth, production and nutrient cycling. Therefore, the breeding program should consider genetics of traits. However, measurement methods could seriously affect progress selection program. In this study, we tested ability spectroscopy quantify specific leaf including anthocyanins (ANTH), flavonoids (FLAV) nitrogen balance index (NBI), estimated genetic variation these based on spectroscopic predicted data. Fresh leaves Sassafras tzumu were selected for spectral collection ANTH, FLAV NBI concentrations by standard analytical methods. Partial least squares regression (PLSR), five spectra pre-processing methods, four variable algorisms conducted optimal model selection. Each trait was simulated 200 times error estimation. Results The normal variate (SNV) ANTH 1st derivatives models, combined with significant Multivariate Correlation (sMC) algorithm are finally regarded as best performance models. produced highest accuracy prediction a mean R 2 0.72 RMSE 0.10%, followed (mean 0.58, 0.11% 0.44, 0.04%). High heritability found h 0.78, 0.58 0.61 respectively. It shows that it is beneficial possible improvement Conclusions Spectroscopy can successfully characterize living simultaneous multiple plant provides promising high-throughput tool quick analysis large size samples serves