A prospective framework for testing whether pairwise transcriptomic module coordination can represent drought-associated root architectural change

Abstract

Soil drying changes root tissue distribution and amount in the soil, while the shoot limits water loss. Root length and surface area describe how extensively roots explore and contact the soil. Root volume and average diameter reflect tissue investment and radial growth. Together, these traits show only the final architectural outcome. They do not tell us whether the change comes from plastic adjustment, slower development, or damage. Bulk-root transcriptomes further complicate interpretation by combining signals from tissues, developmental zones, and cell types. We still do not know whether drought-related changes in root architecture can be represented by relatively simple coordination among transcriptomic modules, or whether more complex combinations are needed. We therefore propose a prospective framework to test whether module eigengenes and their pairwise products adequately predict root architectural change. Root architectural deformation (RAD) is defined as a candidate geometric coordinate from four fixed traits using generalized contrastive principal component analysis (gcPCA) after pooled common standardization. Its variance-ratio objective avoids manual choice of the contrast parameter used in classical cPCA. Testing still depends on Clause Zero: phenotype and transcriptome measurements must be linked to independently measured soil water status and collected within a physiologically appropriate window. Module eigengenes and their pairwise products form the specified predictor basis. Topology-based weighting, locked LASSO, and matched target-decoy screening narrow the candidate set. Clause A asks whether retained terms rise above a structured-permutation ceiling. Clause B asks whether a pre-specified higher-capacity model recovers out-of-sample predictive information beyond this basis, independently of candidate survival. Public wheat phenotypes show cPCA sensitivity and favorable local gcPCA diagnostics, but do not establish a stable plasticity phenotype. Controlled cases illustrate intended readouts; the full empirical framework remains untested. A meaningful gain from the higher-capacity model would argue against pairwise practical adequacy. Failure to find a clear advantage would not, by itself, establish pairwise adequacy. A non-inferiority analysis against a prospectively justified utility margin would still be necessary. Biological interpretation comes later and would depend on independent validation, root-tissue localization, perturbation experiments, and evidence across genotypes, environments, and yield.