Speakers - 2026

Title: On employing deep learning to predict high yield wheat transcriptomes

Abstract

For the industrial sector, crops are the main source of food and raw materials. Plant pests and unfavorable environmental factors constantly disturb the equilibrium between crop yield and food intake. The average yield of wheat has been decreasing overtime since 1980. Therefore there is dire need to ensure global food security by all means. It has been predicted that production of food should increase by approximately 70% for it to be able to feed the rising expected population of over 9billion people. The main reason for the existing poor wheat yields is sub-optimal weed management practices whereby genomic information is required for greater yields of wheat in future. Notably, it is not possible to cover up a complete yield gap either environmentally or economically because of climatic risks, harsh environmental impact and diminishing returns. In this case the best way to enhance yields of a crop like wheat is to identify the genetic yield potential. Here, in this study we performed transcriptomic analysis of wheat. We took dataset of wheat microarray dataset and differentially expressed genes. The cutoff criteria for identifying differentially expressed genes (DEGs) for all the comparisons was p-value <0.05. Different bioinformatics tools used were DAVID for Gene Ontology (GO) terms and pathway analysis and Cytoscape with its different plugins such as stringAPP, MCODE, CytoHubba for construction of network and identification of hub genes and their clusters. We identified significant number of common DEGs of the dataset and their role in different pathways. HSD11Bl, RAB, HIR1 are important and significant genes in wheat yield prediction. Artificial intelligence technique like ANN is performed. An ANN has many artificial neurons which are called units. The input layers have our data which need to be processed. For this the data goes through hidden layers and gives out. In future the identified hub genes have been evaluated by various wet lab techniques and it may be used for potential target for wheat yield prediction.