Abstract:Artificial Intelligence (AI) has emerged as a key driver of precision agriculture, facilitating enhanced crop productivity, optimized resource use, farm sustainability, and informed decision-making. Also, the expansion of genome sequencing technology has greatly increased crop genomic resources, deepening our understanding of genetic variation and enhancing desirable crop traits to optimize performance in various environments. There is increasing interest in using machine learning (ML) and deep learning (DL) algorithms for genotype-to-phenotype prediction due to their excellence in capturing complex interactions within large, high-dimensional datasets. In this work, we propose a new LSTM autoencoder-based model for barley genotype-to-phenotype prediction, specifically for flowering time and grain yield estimation, which could potentially help optimize yields and management practices. Our model outperformed the other baseline methods, demonstrating its potential in handling complex high-dimensional agricultural datasets and enhancing crop phenotype prediction performance.

What problem does this paper attempt to address?

The paper attempts to address the problem of predicting the relationship between barley genotypes and phenotypes using deep neural networks, specifically models based on LSTM autoencoders. The specific applications are in predicting flowering time and grain yield. This research aims to optimize barley yield and management practices, thereby improving the efficiency and sustainability of agricultural production. ### Background and Motivation 1. **Need for Precision Agriculture**: With the growing global population and intensifying climate change, the demand for sustainable agricultural practices is becoming increasingly urgent. Artificial Intelligence (AI), especially Machine Learning (ML) and Deep Learning (DL), shows great potential in enhancing agricultural productivity, optimizing resource utilization, improving farm sustainability, and supporting decision-making. 2. **Growth of Genomic Resources**: The development of genomic sequencing technologies has significantly increased crop genomic resources, deepened the understanding of genetic variation, and enhanced the improvement of crop traits to adapt to various environments. 3. **Importance of Genotype-to-Phenotype Prediction**: Understanding the relationship between genotype and phenotype is crucial for improving crop performance and resilience, ensuring food security, and sustainable development. Traditional statistical methods have limitations in handling high-dimensional data, whereas ML and DL algorithms can capture complex high-order interactions, improving prediction accuracy. ### Research Objectives 1. **Propose a New Model**: This paper proposes a deep neural network model based on LSTM autoencoders for predicting barley genotype-to-phenotype relationships, specifically for flowering time and grain yield prediction. 2. **Optimize Prediction Performance**: By pre-training the LSTM autoencoder, latent feature representations of high-dimensional genomic data are extracted, thereby improving the model's prediction performance when handling complex high-dimensional agricultural datasets. 3. **Validate Model Effectiveness**: Through experiments, the model's performance in predicting barley flowering time and grain yield is validated and compared with baseline methods, demonstrating its superiority. ### Main Contributions 1. **Model Innovation**: Introducing the LSTM autoencoder structure, which extracts latent feature representations of genomic data through pre-training, reducing reliance on traditional feature engineering. 2. **Performance Improvement**: Experimental results show that the model outperforms other baseline methods in predicting barley flowering time and grain yield, demonstrating its potential in handling complex high-dimensional datasets. 3. **Practical Application**: The model is expected to play an important role in optimizing barley yield and management practices, improving the efficiency and sustainability of agricultural production. ### Conclusion The paper successfully proposes and validates a deep neural network model based on LSTM autoencoders for predicting barley genotype-to-phenotype relationships. The model performs excellently in predicting barley flowering time and grain yield, showing high practical value and potential application prospects. Future research will further expand the model's application scope, including the incorporation of different crop types and time-series environmental variables.

LSTM Autoencoder-based Deep Neural Networks for Barley Genotype-to-Phenotype Prediction

Using Multi-Sensor Data Fusion Techniques and Machine Learning Algorithms for Improving UAV-Based Yield Prediction of Oilseed Rape

Automatic and fast classification of barley grains from images: A deep learning approach

Crop yield prediction integrating genotype and weather variables using deep learning

Phenotype prediction and genome-wide association study using deep convolutional neural network of soybean

InceptionV3-LSTM: A Deep Learning Net for the Intelligent Prediction of Rapeseed Harvest Time

Assimilation of Deep Learning and Machine Learning Schemes into a Remote Sensing-Incorporated Crop Model to Simulate Barley and Wheat Productivities

A deep auto-encoder model for gene expression prediction

A deep convolutional neural network approach for predicting phenotypes from genotypes

GPTransformer: A Transformer-Based Deep Learning Method for Predicting Fusarium Related Traits in Barley

Application Of Deep Learning In Genomic Selection

Linking genetic markers and crop model parameters using neural networks to enhance genomic prediction of integrative traits

Deep learning methods improve genomic prediction of wheat breeding

Deep Learning for Predicting Complex Traits in Spring Wheat Breeding Program

A comparative study of 11 non-linear regression models highlighting autoencoder, DBN, and SVR, enhanced by SHAP importance analysis in soybean branching prediction

Genome-Wide Prediction of Complex Traits in Two Outcrossing Plant Species Through Deep Learning and Bayesian Regularized Neural Network

Multimodal deep learning methods enhance genomic prediction of wheat breeding

A New Framework for Winter Wheat Yield Prediction Integrating Deep Learning and Bayesian Optimization

Plant Genotype to Phenotype Prediction Using Machine Learning

Classification and prediction of drought and salinity stress tolerance in barley using GenPhenML

A New Deep Learning Calibration Method Enhances Genome-Based Prediction of Continuous Crop Traits