Additional file 4 of Smoothing and extraction of traits in the growth analysis of noninvasive phenotypic data

Additional file 4.R scripts and data for preparing the tomato data and carrying out the reported analyses. The data is provided in the file tomato.dat.csv, but in R is also available with the growthPheno package. The script global.r provides settings, constants and functions that are used across all scripts and is executed at the beginning of most scripts. The script SET.r gives the code for obtaining the smoothed longitudinal data (Steps 1–4 of the SET process). Cart.dat.r extracts the per-cart.traits (Step 5 of the SET process). Cart.anal.r analyses the per-cart data and Cart.predict.r obtains the predictions based on the selected models (Step 6 of the SET-based analysis). Cart.joint.r performs the extra joint analysis of per-cart traits. Longi.anal.r fits several models to all the tomato data for PSA and ln(PSA) in order to establish a variance model for each and then, for the selected variance model, the number of knots for the splines describing the curved trend for each combination of Zn and AMF is varied (Stages 1–2). Longi.predict.r obtains the predictions for the different numbers of knots (Stage 3). Longi.trend.r investigates the effect of Zn and AMF on the time trend when 10 knots are used and does diagnostic checking of the residuals (Stage 4); it also fits a reduced variance model that assumes equal variances for different DAPs and zero correlation between DAPs.