Code and data for: Layered predictive breeding: when do omics layers earn their place in crop genomic prediction?

Code and data for Fig. 5, Supplementary Fig. S1 and Supplementary Table S3 of the review "Layered predictive breeding: when do omics layers earn their place in crop genomic prediction?" (Anas et al., submitted to Plant Communications). Fig5_multistudy_data_v3.csv: 270 rows, one per study x trait x omics layer x predictor type. Each value is the prediction value printed in the source study (correlation r in most rows; column "metric" marks the exceptions), with its page or table in column "source"; the same rows appear in Supplementary Table S1. Fig5_summary.R draws Fig. 5 (study-level summary: median and range of the gain over the genomic baseline). SuppFigS1.R draws Supplementary Fig. S1 (every trait row). Study_level_analysis.py produces the study-level summary and sensitivity analyses in Supplementary Table S3 (median per study, bootstrap 95% interval, Wilcoxon test, size-weighted median, leave-one-study-out, Spearman small-study check). Gain (column "delta") = value for the omics or genomic + omics predictor minus value for the genomic baseline, within a study, on the same lines and under the same cross-validation scheme. No raw data from the source studies were re-analysed; the scripts summarize the printed values.