Abstract: One or more fields can be divided into portions, and different sets of interventions can be performed during the course of planting and growing a crop in each field portion. The effect of each set of interventions on the crop outcome of the corresponding field portion can be determined using a best performing model of a set of predictive models applied to the set of interventions. The selected model can be used to determine an effect of each intervention on crop outcome of a field portion by estimating a portion of the crop outcome attributable to the intervention. A machine-learned model can be trained based on the effect each intervention has on crop outcome determined by the selected model, and the machine-learned model can be applied to subsequent fields and crops to predict effects that various interventions can have on subsequent crop outcomes.

Abstract: This invention relates to methods and compositions for improving plant health. The present invention includes methods for making an enriched library of treatments capable of improving plant health, methods for an making an enriched library of plants capable of being improved by a treatment, and methods of marketing a plant treatment, as well as synthetic compositions comprising treatments produced by the methods of the present invention, synthetic compositions comprising endophytes capable of improving plant health, and nucleic acid probes and nucleic acid detection kits that may be used in the methods of the present invention.