PlantDeepSEA, a deep learning-based web service to predict the regulatory effects of genomic variants in plants

Hu Zhao,Zhuo Tu,Yinmeng Liu,Zhanxiang Zong,Jiacheng Li,Hao Liu,Feng Xiong,Jinling Zhan,Xuehai Hu,Weibo Xie

DOI: https://doi.org/10.1093/nar/gkab383

IF: 14.9

2021-05-25

Nucleic Acids Research

Abstract:Abstract Characterizing regulatory effects of genomic variants in plants remains a challenge. Although several tools based on deep-learning models and large-scale chromatin-profiling data have been available to predict regulatory elements and variant effects, no dedicated tools or web services have been reported in plants. Here, we present PlantDeepSEA as a deep learning-based web service to predict regulatory effects of genomic variants in multiple tissues of six plant species (including four crops). PlantDeepSEA provides two main functions. One is called Variant Effector, which aims to predict the effects of sequence variants on chromatin accessibility. Another is Sequence Profiler, a utility that performs ‘in silico saturated mutagenesis’ analysis to discover high-impact sites (e.g., cis-regulatory elements) within a sequence. When validated on independent test sets, the area under receiver operating characteristic curve of deep learning models in PlantDeepSEA ranges from 0.93 to 0.99. We demonstrate the usability of the web service with two examples. PlantDeepSEA could help to prioritize regulatory causal variants and might improve our understanding of their mechanisms of action in different tissues in plants. PlantDeepSEA is available at http://plantdeepsea.ncpgr.cn/.

biochemistry & molecular biology

What problem does this paper attempt to address?

The paper aims to address the challenging problem of predicting the regulatory effects of genomic variations in plants. Specifically, although there are some tools based on deep learning models and large-scale chromatin mapping data available for predicting regulatory elements and variant effects, there are no dedicated tools or web services for this purpose in plants. Therefore, the authors developed PlantDeepSEA, a deep learning-based web service platform for predicting the regulatory effects of genomic variations in multiple tissues of six plant species (including four crops). PlantDeepSEA offers two main functionalities: 1. **Variant Effector**: Predicts the impact of sequence variations on chromatin accessibility. 2. **Sequence Profiler**: Performs "in silico saturation mutagenesis" analysis to identify high-impact sites in the sequence (e.g., cis-regulatory elements). Through validation on independent test sets, the deep learning models in PlantDeepSEA achieved an area under the receiver operating characteristic curve (AUROC) ranging from 0.93 to 0.99. This indicates that the platform has high predictive accuracy and can help prioritize regulatory causal variants, thereby improving our understanding of plant regulatory mechanisms in different tissues.

PlantDeepSEA, a deep learning-based web service to predict the regulatory effects of genomic variants in plants

ChIP-Hub Provides an Integrative Platform for Exploring Plant Regulome

A Hybrid Prediction Method for Plant lncRNA-Protein Interaction.

PmliPred: a method based on hybrid model and fuzzy decision for plant miRNA-lncRNA interaction prediction.

A deep learning approach to automate whole‐genome prediction of diverse epigenomic modifications in plants

Modeling 0.6 million genes for the rational design of functional cis-regulatory variants and de novo design of cis- regulatory sequences

A Deep Learning-Based Sequence Analyzer Incorporating the Transcription Factor Binding Affinity to Dissect the Effects of Non-Coding Genetic Variants

PPVED: A machine learning tool for predicting the effect of single amino acid substitution on protein function in plants

Deep learning facilitates precise identification of disease-resistance genes in plants

Deep6mAPred: A CNN and Bi-LSTM-based deep learning method for predicting DNA N6-methyladenosine sites across plant species

Predicting Functional Elements and Variants Effects in Non-Coding Regions Based on Deep Learning

Cross-species modeling of plant genomes at single nucleotide resolution using a pre-trained DNA language model

Models trained to predict differential expression across plant organs identify distal and proximal regulatory regions

PlantC2U: Deep learning of cross-species sequence landscapes predict plastid C-to-U RNA editing in plants

Prediction of condition-specific regulatory genes using machine learning

Deep learning approaches for non-coding genetic variant effect prediction: current progress and future prospects

Identification of Novel Regulators of Leaf Senescence Using a Deep Learning Model

Machine learning: A powerful tool for gene function prediction in plants

DeepGenGrep: a general deep learning-based predictor for multiple genomic signals and regions

agReg-SNPdb-Plants: A Database of Regulatory SNPs for Agricultural Plant Species