McLean, G. | bioRxiv Science

bioRxiv2023

New algorithm for pearl millet modelling in APSIM allowing a mechanistic simulation of tillers

We present a new algorithm for pearl millet simulation in APSIM. Compared to the actual released model, this new model increases the ability to simulate dynamic tillers by integrating recent progresses about biological understanding of the tillering mechanism. The new algorithm also offers the possibility to have an increased genetic control over key functions like canopy development and tillering through additional genotype related parameters. Next to model description, we also present the parametrization of 9444 and HHB 67-2, two genotypes broadly used in India. Overall, we could show that the new algorithm is able to reconstruct the main plant function like biomass accumulation and tillering. Some margin of improvement remains concerning the simulation of tiller cessation.

plant biology↗

bioRxiv2021

Predicting phenotypes from genetic, environment, management, and historical data using CNNs

Predicting phenotypes from genetic (G), environmental (E), and management (M) conditions is a long-standing challenge with implications to agriculture, medicine, and conservation. Most methods reduce the factors in a dataset (feature engineering) in a subjective and potentially oversimplified manner. Convolutional Neural Networks (CNN) can overcome this by allowing the data itself to determine which factors are most important. CNN models were developed for predicting agronomic yield from a combination of replicated trials and historical yield survey data. The results were more accurate than standard methods when tested on heldout G, E, and M data (r=0.5 vs r=0.4), and performed slightly worse than standard methods when only G was held out (r=0.74 vs r=0.78). Pre-training on historical data increased accuracy by 1-36% compared to trial data alone. Saliency map analysis indicated the CNN has "learned" to prioritize many factors of known agricultural importance.

plant biology↗