Inference of Transcription Regulatory Network in Low Phytic Acid Soybean Seeds

A review of artificial intelligence-assisted omics techniques in plant defense: current trends and future directions

Plants intricately deploy defense systems to counter diverse biotic and abiotic stresses. Omics technologies, spanning genomics, transcriptomics, proteomics, and metabolomics, have revolutionized the exploration of plant defense mechanisms, unraveling molecular intricacies in response to various stressors. However, the complexity and scale of omics data necessitate sophisticated analytical tools for meaningful insights. This review delves into the application of artificial intelligence algorithms, particularly machine learning and deep learning, as promising approaches for deciphering complex omics data in plant defense research. The overview encompasses key omics techniques and addresses the challenges and limitations inherent in current AI-assisted omics approaches. Moreover, it contemplates potential future directions in this dynamic field. In summary, AI-assisted omics techniques present a robust toolkit, enabling a profound understanding of the molecular foundations of plant defense and paving the way for more effective crop protection strategies amidst climate change and emerging diseases.

Machine learning bridges omics sciences and plant breeding

Some of the biological knowledge obtained from fundamental research will be implemented in applied plant breeding. To bridge basic research and breeding practice, machine learning (ML) holds great promise to translate biological knowledge and omics data into precision-designed plant breeding. Here, we review ML for multi-omics analysis in plants, including data dimensionality reduction, inference of gene-regulation networks, and gene discovery and prioritization. These applications will facilitate understanding trait regulation mechanisms and identifying target genes potentially applicable to knowledge-driven molecular design breeding. We also highlight applications of deep learning in plant phenomics and ML in genomic selection-assisted breeding, such as various ML algorithms that model the correlations among genotypes (genes), phenotypes (traits), and environments, to ultimately achieve data-driven genomic design breeding.

A comprehensive transcription factor and DNA-binding motif resource for the construction of gene regulatory networks in Botrytis cinerea and Trichoderma atroviride

Botrytis cinerea and Trichoderma atroviride are two relevant fungi in agricultural systems. To gain insights into these organisms' transcriptional gene regulatory networks (GRNs), we generated a manually curated transcription factor (TF) dataset for each of them, followed by a GRN inference utilizing available sequence motifs describing DNA-binding specificity and global gene expression data. As a proof of concept of the usefulness of this resource to pinpoint key transcriptional regulators, we employed publicly available transcriptomics data and a newly generated dual RNA-seq dataset to build context-specific Botrytis and Trichoderma GRNs under two different biological paradigms: exposure to continuous light and Botrytis-Trichoderma confrontation assays. Network analysis of fungal responses to constant light revealed striking differences in the transcriptional landscape of both fungi. On the other hand, we found that the confrontation of both microorganisms elicited a distinct set of differentially expressed genes with changes in T. atroviride exceeding those in B. cinerea. Using our regulatory network data, we were able to determine, in both fungi, central TFs involved in this interaction response, including TFs controlling a large set of extracellular peptidases in the biocontrol agent T. atroviride. In summary, our work provides a comprehensive catalog of transcription factors and regulatory interactions for both organisms. This catalog can now serve as a basis for generating novel hypotheses on transcriptional regulatory circuits in different experimental contexts.

Network Inference of Transcriptional Regulation in Germinating Low Phytic Acid Soybean Seeds

The low phytic acid (lpa) trait in soybeans can be conferred by loss-of-function mutations in genes encoding myo-inositol phosphate synthase and two epistatically interacting genes encoding multidrug-resistance protein ATP-binding cassette (ABC) transporters. However, perturbations in phytic acid biosynthesis are associated with poor seed vigor. Since the benefits of the lpa trait, in terms of end-use quality and sustainability, far outweigh the negatives associated with poor seed performance, a fuller understanding of the molecular basis behind the negatives will assist crop breeders and engineers in producing variates with lpa and better germination rate. The gene regulatory network (GRN) for developing low and normal phytic acid soybean seeds was previously constructed, with genes modulating a variety of processes pertinent to phytic acid metabolism and seed viability being identified. In this study, a comparative time series analysis of low and normal phytic acid soybeans was carried out to investigate the transcriptional regulatory elements governing the transitional dynamics from dry seed to germinated seed. GRNs were reverse engineered from time series transcriptomic data of three distinct genotypic subsets composed of lpa soybean lines and their normal phytic acid sibling lines. Using a robust unsupervised network inference scheme, putative regulatory interactions were inferred for each subset of genotypes. These interactions were further validated by published regulatory interactions found in Arabidopsis thaliana and motif sequence analysis. Results indicate that lpa seeds have increased sensitivity to stress, which could be due to changes in phytic acid levels, disrupted inositol phosphate signaling, disrupted phosphate ion (Pi) homeostasis, and altered myo-inositol metabolism. Putative regulatory interactions were identified for the latter two processes. Changes in abscisic acid (ABA) signaling candidate transcription factors (TFs) putatively regulating genes in this process were identified as well. Analysis of the GRNs reveal altered regulation in processes that may be affecting the germination of lpa soybean seeds. Therefore, this work contributes to the ongoing effort to elucidate molecular mechanisms underlying altered seed viability, germination and field emergence of lpa crops, understanding of which is necessary in order to mitigate these problems.

Using Network-Based Machine Learning to Predict Transcription Factors Involved in Drought Resistance

Gene regulatory networks underpin stress response pathways in plants. However, parsing these networks to prioritize key genes underlying a particular trait is challenging. Here, we have built the Gene Regulation and Association Network (GRAiN) of rice (Oryza sativa). GRAiN is an interactive query-based web-platform that allows users to study functional relationships between transcription factors (TFs) and genetic modules underlying abiotic-stress responses. We built GRAiN by applying a combination of different network inference algorithms to publicly available gene expression data. We propose a supervised machine learning framework that complements GRAiN in prioritizing genes that regulate stress signal transduction and modulate gene expression under drought conditions. Our framework converts intricate network connectivity patterns of 2160 TFs into a single drought score. We observed that TFs with the highest drought scores define the functional, structural, and evolutionary characteristics of drought resistance in rice. Our approach accurately predicted the function of OsbHLH148 TF, which we validated using in vitro protein-DNA binding assays and mRNA sequencing loss-of-function mutants grown under control and drought stress conditions. Our network and the complementary machine learning strategy lends itself to predicting key regulatory genes underlying other agricultural traits and will assist in the genetic engineering of desirable rice varieties.

Identification of Gene Regulatory Networks from Single-Cell Expression Data

Single-cell RNAseq is an emerging technology that allows the quantification of gene expression in individual cells. In plants, single-cell sequencing technology has been applied to generate root cell expression maps under many experimental conditions. DAP-seq and ATAC-seq have also been used to generate genome-scale maps of protein-DNA interactions and open chromatin regions in plants. In this protocol, we describe a multistep computational pipeline for the integration of single-cell RNAseq data with DAP-seq and ATAC-seq data to predict regulatory networks and key regulatory genes. Our approach utilizes machine learning methods including feature selection and stability selection to identify candidate regulatory genes. The network generated by this pipeline can be used to provide a putative annotation of gene regulatory modules and to identify candidate transcription factors that could play a key role in specific cell types.

Using Network-Based Machine Learning to Predict Transcription Factors Involved in Drought Resistance

Gene regulatory networks underpin stress response pathways in plants. However, parsing these networks to prioritize key genes underlying a particular trait is challenging. Here, we have built the Gene Regulation and Association Network (GRAiN) of rice (Oryza sativa). GRAiN is an interactive query-based web-platform that allows users to study functional relationships between transcription factors (TFs) and genetic modules underlying abiotic-stress responses. We built GRAiN by applying a combination of different network inference algorithms to publicly available gene expression data. We propose a supervised machine learning framework that complements GRAiN in prioritizing genes that regulate stress signal transduction and modulate gene expression under drought conditions. Our framework converts intricate network connectivity patterns of 2160 TFs into a single drought score. We observed that TFs with the highest drought scores define the functional, structural, and evolutionary characteristics of drought resistance in rice. Our approach accurately predicted the function of OsbHLH148 TF, which we validated using in vitro protein-DNA binding assays and mRNA sequencing loss-of-function mutants grown under control and drought stress conditions. Our network and the complementary machine learning strategy lends itself to predicting key regulatory genes underlying other agricultural traits and will assist in the genetic engineering of desirable rice varieties.

Genetic interactions regulating seed phytate and oligosaccharides in soybean (Glycine max L.)

Two low-phytate soybean (Glycine max (L.) Merr.) mutant lines- V99-5089 (mips mutation on chromosome 11) and CX-1834 (mrp-l and mrp-n mutations on chromosomes 19 and 3, respectively) have proven to be valuable resources for breeding of low-phytate, high-sucrose, and low-raffinosaccharide soybeans, traits that are highly desirable from a nutritional and environmental standpoint. A recombinant inbred population derived from the cross CX1834 x V99-5089 provides an opportunity to study the effect of different combinations of these three mutations on soybean phytate and oligosaccharides levels. Of the 173 recombinant inbred lines tested, 163 lines were homozygous for various combinations of MIPS and two MRP loci alleles. These individuals were grouped into eight genotypic classes based on the combination of SNP alleles at the three mutant loci. The two genotypic classes that were homozygous mrp-l/mrp-n and either homozygous wild-type or mutant at the mips locus (MIPS/mrp-l/mrp-n or mips/mrp-l/mrp-n) displayed relatively similar ~55% reductions in seed phytate, 6.94 mg g -1 and 6.70 mg g-1 respectively, as compared with 15.2 mg g-1 in the wild-type MIPS/MRP-L/MRP-N seed. Therefore, in the presence of the double mutant mrp-l/mrp-n, the mips mutation did not cause a substantially greater decrease in seed phytate level. However, the nutritionally-desirable high-sucrose/low-stachyose/low-raffinose seed phenotype originally observed in soybeans homozygous for the mips allele was reversed in the presence of mrp-l/mrp-n mutations: homozygous mips/mrp-l/mrp-n seed displayed low-sucrose (7.70%), high-stachyose (4.18%), and the highest observed raffinose (0.94%) contents per gram of dry seed. Perhaps the block in phytic acid transport from its cytoplasmic synthesis site to its storage site, conditioned by mrp-l/mrp-n, alters myo-inositol flux in mips seeds in a way that restores to wild-type levels the mips conditioned reductions in raffinosaccharides. Overall this study determined the combinatorial effects of three low phytic acid causing mutations on regulation of seed phytate and oligosaccharides in soybean.

Prediction of condition-specific regulatory genes using machine learning

Recent advances in genomic technologies have generated data on large-scale protein-DNA interactions and open chromatin regions for many eukaryotic species. How to identify condition-specific functions of transcription factors using these data has become a major challenge in genomic research. To solve this problem, we have developed a method called ConSReg, which provides a novel approach to integrate regulatory genomic data into predictive machine learning models of key regulatory genes. Using Arabidopsis as a model system, we tested our approach to identify regulatory genes in data sets from single cell gene expression and from abiotic stress treatments. Our results showed that ConSReg accurately predicted transcription factors that regulate differentially expressed genes with an average auROC of 0.84, which is 23.5-25% better than enrichment-based approaches. To further validate the performance of ConSReg, we analyzed an independent data set related to plant nitrogen responses. ConSReg provided better rankings of the correct transcription factors in 61.7% of cases, which is three times better than other plant tools. We applied ConSReg to Arabidopsis single cell RNA-seq data, successfully identifying candidate regulatory genes that control cell wall formation. Our methods provide a new approach to define candidate regulatory genes using integrated genomic data in plants.

Gene Regulatory Network Inference: Connecting Plant Biology and Mathematical Modeling

Plant responses to environmental and intrinsic signals are tightly controlled by multiple transcription factors (TFs). These TFs and their regulatory connections form gene regulatory networks (GRNs), which provide a blueprint of the transcriptional regulations underlying plant development and environmental responses. This review provides examples of experimental methodologies commonly used to identify regulatory interactions and generate GRNs. Additionally, this review describes network inference techniques that leverage gene expression data to predict regulatory interactions. These computational and experimental methodologies yield complex networks that can identify new regulatory interactions, driving novel hypotheses. Biological properties that contribute to the complexity of GRNs are also described in this review. These include network topology, network size, transient binding of TFs to DNA, and competition between multiple upstream regulators. Finally, this review highlights the potential of machine learning approaches to leverage gene expression data to predict phenotypic outputs.

A transcriptional regulatory network of Rsv3-mediated extreme resistance against Soybean mosaic virus

Resistance genes are an effective means for disease control in plants. They predominantly function by inducing a hypersensitive reaction, which results in localized cell death restricting pathogen spread. Some resistance genes elicit an atypical response, termed extreme resistance, where resistance is not associated with a hypersensitive reaction and its standard defense responses. Unlike hypersensitive reaction, the molecular regulatory mechanism(s) underlying extreme resistance is largely unexplored. One of the few known, naturally occurring, instances of extreme resistance is resistance derived from the soybean Rsv3 gene, which confers resistance against the most virulent Soybean mosaic virus strains. To discern the regulatory mechanism underlying Rsv3-mediated extreme resistance, we generated a gene regulatory network using transcriptomic data from time course comparisons of Soybean mosaic virus-G7-inoculated resistant (L29, Rsv3-genotype) and susceptible (Williams82, rsv3-genotype) soybean cultivars. Our results show Rsv3 begins mounting a defense by 6 hpi via a complex phytohormone network, where abscisic acid, cytokinin, jasmonic acid, and salicylic acid pathways are suppressed. We identified putative regulatory interactions between transcription factors and genes in phytohormone regulatory pathways, which is consistent with the demonstrated involvement of these pathways in Rsv3-mediated resistance. One such transcription factor identified as a putative transcriptional regulator was MYC2 encoded by Glyma.07G051500. Known as a master regulator of abscisic acid and jasmonic acid signaling, MYC2 specifically recognizes the G-box motif ("CACGTG"), which was significantly enriched in our data among differentially expressed genes implicated in abscisic acid- and jasmonic acid-related activities. This suggests an important role for Glyma.07G051500 in abscisic acid- and jasmonic acid-derived defense signaling in Rsv3. Resultantly, the findings from our network offer insights into genes and biological pathways underlying the molecular defense mechanism of Rsv3-mediated extreme resistance against Soybean mosaic virus. The computational pipeline used to reconstruct the gene regulatory network in this study is freely available at https://github.com/LiLabAtVT/rsv3-network.

Computational prediction of gene regulatory networks in plant growth and development

Plants integrate a wide range of cellular, developmental, and environmental signals to regulate complex patterns of gene expression. Recent advances in genomic technologies enable differential gene expression analysis at a systems level, allowing for improved inference of the network of regulatory interactions between genes. These gene regulatory networks, or GRNs, are used to visualize the causal regulatory relationships between regulators and their downstream target genes. Accordingly, these GRNs can represent spatial, temporal, and/or environmental regulations and can identify functional genes. This review summarizes recent computational approaches applied to different types of gene expression data to infer GRNs in the context of plant growth and development. Three stages of GRN inference are described: first, data collection and analysis based on the dataset type; second, network inference application based on data availability and proposed hypotheses; and third, validation based on in silico, in vivo, and in planta methods. In addition, this review relates data collection strategies to biological questions, organizes inference algorithms based on statistical methods and data types, discusses experimental design considerations, and provides guidelines for GRN inference with an emphasis on the benefits of integrative approaches, especially when a priori information is limited. Finally, this review concludes that computational frameworks integrating large-scale heterogeneous datasets are needed for a more accurate (e.g. fewer false interactions), detailed (e.g. discrimination between direct versus indirect interactions), and comprehensive (e.g. genetic regulation under various conditions and spatial locations) inference of GRNs.

Statistical and Machine Learning Approaches to Predict Gene Regulatory Networks From Transcriptome Datasets

Statistical and machine learning (ML)-based methods have recently advanced in construction of gene regulatory network (GRNs) based on high-throughput biological datasets. GRNs underlie almost all cellular phenomena; hence, comprehensive GRN maps are essential tools to elucidate gene function, thereby facilitating the identification and prioritization of candidate genes for functional analysis. High-throughput gene expression datasets have yielded various statistical and ML-based algorithms to infer causal relationship between genes and decipher GRNs. This review summarizes the recent advancements in the computational inference of GRNs, based on large-scale transcriptome sequencing datasets of model plants and crops. We highlight strategies to select contextual genes for GRN inference, and statistical and ML-based methods for inferring GRNs based on transcriptome datasets from plants. Furthermore, we discuss the challenges and opportunities for the elucidation of GRNs based on large-scale datasets obtained from emerging transcriptomic applications, such as from population-scale, single-cell level, and life-course transcriptome analyses.