Classification and Visualization of Genotype x Phenotype Interactions in Biomass Sorghum
This work addresses the challenge of uncovering genetic-phenotypic relationships in crops for agricultural researchers, but it is incremental as it applies existing CNN and visualization techniques to a new dataset.
The authors tackled the problem of understanding genotype-phenotype interactions in biomass sorghum by training deep CNNs to classify plant images based on SNPs and using visualizations to highlight key features, demonstrating the method's capacity with RGB imagery from the TERRA-REF gantry.
We introduce a simple approach to understanding the relationship between single nucleotide polymorphisms (SNPs), or groups of related SNPs, and the phenotypes they control. The pipeline involves training deep convolutional neural networks (CNNs) to differentiate between images of plants with reference and alternate versions of various SNPs, and then using visualization approaches to highlight what the classification networks key on. We demonstrate the capacity of deep CNNs at performing this classification task, and show the utility of these visualizations on RGB imagery of biomass sorghum captured by the TERRA-REF gantry. We focus on several different genetic markers with known phenotypic expression, and discuss the possibilities of using this approach to uncover genotype x phenotype relationships.