Search

Plants or tissues can be regenerated through various pathways. Like animal regeneration, cell totipotency and pluripotency are the molecular basis of plant regeneration. Detailed systematic studies on Arabidopsis thalianagradually unravel the fundamental mechanisms and principles underlying plant regeneration. Specifically, plant hormones, cell division, epigenetic remodeling, and transcription factors play crucial roles in reprogramming somatic cells and reestablishing meristematic cells. Recent research on basal non-vascular plants and monocot crops has revealed that plant regeneration differs among species, with various plant species using distinct mechanisms and displaying significant differences in regenerative capacity. Conducting multi-omics studies at the single-cell level, tracking plant regeneration processes in real-time, and deciphering the natural variation in regenerative capacity will ultimately help understand the essence of plant regeneration, improve crop regeneration efficiency, and contribute to future crop design.

The goal of supervised representation learning is to construct effective data representations for prediction. Among all the characteristics of an ideal nonparametric representation of high-dimensional complex data, sufficiency, low dimensionality and disentanglement are some of the most essential ones. We propose a deep dimension reduction approach to learning representations with these characteristics. The proposed approach is a nonparametric generalization of the sufficient dimension reduction method. We formulate the ideal representation learning task as that of finding a nonparametric representation that minimizes an objective function characterizing conditional independence and promoting disentanglement at the population level. We then estimate the target representation at the sample level nonparametrically using deep neural networks. We show that the estimated deep nonparametric representation is consistent in the sense that its excess risk converges to zero. Our extensive numerical experiments using simulated and real benchmark data demonstrate that the proposed methods have better performance than several existing dimension reduction methods and the standard deep learning models in the context of classification and regression.

The model plant Physcomitrium patenshas played a pivotal role in enhancing our comprehension of plant evolution and development. However, the current genome harbours numerous regions that remain unfinished and erroneous. To address these issues, we generated an assembly using Oxford Nanopore reads and Hi-C mapping. The assembly incorporates telomeric and centromeric regions, thereby establishing it as a near telomere-to-telomere genome except a region in chromosome 1 that is not fully assembled due to its highly repetitive nature. This near telomere-to-telomere genome resolves the chromosome number at 26 and provides a gap-free genome assembly as well as updated gene models to aid future studies using this model organism.

Rapid advances in DNA synthesis techniques have enabled the assembly and engineering of viral and microbial genomes, presenting new opportunities for synthetic genomics in multicellular eukaryotic organisms. These organisms, characterized by larger genomes, abundant transposons and extensive epigenetic regulation, pose unique challenges. Here we report the in vivo assembly of chromosomal fragments in the moss Physcomitrium patens, producing phenotypically virtually wild-type lines in which one-third of the coding region of a chromosomal arm is replaced by redesigned, chemically synthesized fragments. By eliminating 55.8% of a 155 kb endogenous chromosomal region, we substantially simplified the genome without discernible phenotypic effects, implying that many transposable elements may minimally impact growth. We also introduced other sequence modifications, such as PCRTag incorporation, gene locus swapping and stop codon substitution. Despite these substantial changes, the complex epigenetic landscape was normally established, albeit with some three-dimensional conformation alterations. The synthesis of a partial multicellular eukaryotic chromosome arm lays the foundation for the synthetic moss genome project (SynMoss) and paves the way for genome synthesis in multicellular organisms.

The de novo synthesis of genomes has made unprecedented progress and achieved milestones, particularly in bacteria and yeast. However, the process of synthesizing a multicellular plant genome has not progressed at the same pace, due to the complexity of multicellular plant genomes, technical difficulties associated with large genome size and structure, and the intricacies of gene regulation and expression in plants. Here we outline the bottom-up design principles for the de novo synthesis of the Physcomitrium patens(that is, earthmoss) genome. To facilitate international collaboration and accessibility, we have developed and launched a public online design platform called GenoDesigner. This platform offers an intuitive graphical interface enabling users to efficiently manipulate extensive genome sequences, even up to the gigabase level. This tool is poised to greatly expedite the synthesis of the P. patensgenome, offering an essential reference and roadmap for the synthesis of plant genomes.

Crop breeding heavily relies on natural genetic variation. However, additional new variations are desired to meet the increasing human demand. Inflorescence architecture determines grain number per spike, a major determinant of bread wheat (Triticum aestivumL.) yield. Here, using Brachypodium distachyonas a wheat proxy, we identified DUO-B1, encoding an APETALA2/ethylene response factor (AP2/ERF) transcription factor, regulating spike inflorescence architecture in bread wheat. Mutations of DUO-B1lead to mild supernumerary spikelets, increased grain number per spike and, importantly, increased yield under field conditions without affecting other major agronomic traits. DUO-B1suppresses cell division and promotes the expression of BHt/WFZP, whose mutations could lead to branched ‘miracle-wheat’. Pan-genome analysis indicated that DUO-B1has not been utilized in breeding, and holds promise to increase wheat yield further.

Deep neural networks have demonstrated remarkable efficacy in diverse classification tasks. In this paper, we specifically focus on the predictive performance in deep binary classification problems with the sigmoid loss. Given that sigmoid loss is categorized as a non-convex and bounded loss function, it exhibits potential resilience against the disruptive impact of outlier noises. We first derive the convergence rate of the excess misclassification risk for deep ReLU neural networks with the sigmoid loss, a result that attains minimax optimality. To the best of our acknowledge, we are the first to derive the convergence rate for the sigmoid loss. Moreover, we extend our analysis to derive a faster convergence rate under margin assumptions. This achievement renders our findings comparable to those of commonly employed convex loss functions operating under analogous assumptions. Lastly, we undertake a comprehensive validation of the robustness inherent in the sigmoid loss across diverse datasets. [ABSTRACT FROM AUTHOR]

In multicellular and even single-celled organisms, individual components are interconnected at multiscale levels to produce enormously complex biological networks that help these systems maintain homeostasis for development and environmental adaptation. Systems biology studies initially adopted network analysis to explore how relationships between individual components give rise to complex biological processes. Network analysis has been applied to dissect the complex connectivity of mammalian brains across different scales in time and space in The Human Brain Project. In plant science, network analysis has similarly been applied to study the connectivity of plant components at the molecular, subcellular, cellular, organic, and organism levels. Analysis of these multiscale networks contributes to our understanding of how genotype determines phenotype. In this review, we summarized the theoretical framework of plant multiscale networks and introduced studies investigating plant networks by various experimental and computational modalities. We next discussed the currently available analytic methodologies and multi-level imaging techniques used to map multiscale networks in plants. Finally, we highlighted some of the technical challenges and key questions remaining to be addressed in this emerging field.

Bread wheat expanded its habitat from a core area of the Fertile Crescent to global environments within ~10,000 years. The genetic mechanisms of this remarkable evolutionary success are not well understood. By whole-genome sequencing of populations from 25 subspecies within the genera Triticumand Aegilops, we identified composite introgression from wild populations contributing to a substantial portion (4–32%) of the bread wheat genome, which increased the genetic diversity of bread wheat and allowed its divergent adaptation. Meanwhile, convergent adaptation to human selection showed 2- to 16-fold enrichment relative to random expectation—a certain set of genes were repeatedly selected in Triticumspecies despite their drastic differences in ploidy levels and growing zones, indicating the important role of evolutionary constraints in shaping the adaptive landscape of bread wheat. These results showed the genetic necessities of wheat as a global crop and provided new perspectives on transferring adaptive success across species for crop improvement.

The smoothly clipped absolute deviation (SCAD) and the minimax concave penalty (MCP)-penalized regression models are two important and widely used nonconvex sparse learning tools that can handle variable selection and parameter estimation simultaneously and thus have potential applications in various fields, such as mining biological data in high-throughput biomedical studies. Theoretically, these two models enjoy the oracle property even in the high-dimensional settings, where the number of predictors $p$ may be much larger than the number of observations $n$ . However, numerically, it is quite challenging to develop fast and stable algorithms due to their nonconvexity and nonsmoothness. In this article, we develop a fast algorithm for SCAD- and MCP-penalized learning problems. First, we show that the global minimizers of both models are roots of the nonsmooth equations. Then, a semismooth Newton (SSN) algorithm is employed to solve the equations. We prove that the SSN algorithm converges locally and superlinearly to the Karush–Kuhn–Tucker (KKT) points. The computational complexity analysis shows that the cost of the SSN algorithm per iteration is $O(np)$ . Combined with the warm-start technique, the SSN algorithm can be very efficient and accurate. Simulation studies and a real data example suggest that our SSN algorithm, with comparable solution accuracy with the coordinate descent (CD) and the difference of convex (DC) proximal Newton algorithms, is more computationally efficient.

The seamless-L0(SELO) penalty is a smooth function that very closely resembles the L0 penalty, which has been demonstrated theoretically and practically to be effective in nonconvex penalization for variable selection. In this paper, the authors first generalize the SELO penalty to a class of penalties retaining good features of SELO, and then develop variable selection and parameter estimation in Cox models using the proposed generalized SELO (GSELO) penalized log partial likelihood (PPL) approach. The authors show that the GSELO-PPL procedure possesses the oracle property with a diverging number of predictors under certain mild, interpretable regularity conditions. The entire path of GSELO-PPL estimates can be efficiently computed through a smoothing quasi-Newton (SQN) with continuation algorithm. The authors propose a consistent modified BIC (MBIC) tuning parameter selector for GSELO-PPL, and show that under some regularity conditions, the GSELOPPL- MBIC procedure consistently identifies the true model. Simulation studies and real data analysis are conducted to evaluate the finite sample performance of the proposed method.

A better understanding of wheat functional genomics can improve targeted breeding for better agronomic traits and environmental adaptation. However, the lack of gene-indexed mutants and the low transformation efficiency of wheat limit in-depth gene functional studies and genetic manipulation for breeding. In this study, we created a library for KN9204, a popular wheat variety in northern China, with a reference genome, transcriptome, and epigenome of different tissues, using ethyl methyl sulfonate (EMS) mutagenesis. This library contains a vast developmental diversity of critical tissues and transition stages. Exome capture sequencing of 2090 mutant lines using KN9204 genome–designed probes revealed that 98.79% of coding genes had mutations, and each line had an average of 1383 EMS-type SNPs. We identified new allelic variations for crucial agronomic trait-related genes such as Rht-D1, Q, TaTB1, and WFZP. We tested 100 lines with severe mutations in 80 NAC transcription factors (TFs) under drought and salinity stress and identified 13 lines with altered sensitivity. Further analysis of three lines using transcriptome and chromatin accessibility data revealed hundreds of direct NAC targets with altered transcription patterns under salt or drought stress, including SNAC1, DREB2B, CML16, and ZFP182, factors known to respond to abiotic stress. Thus, we have generated and indexed a KN9204 EMS mutant library that can facilitate functional genomics research and offer resources for genetic manipulation of wheat.

Live imaging through confocal laser scanning microscopy enables the recording, analysis, and comparison of the dynamics of shapes and gene expression patterns of plant shoot apical meristems (SAMs) or primordia. Here, we provide a protocol to describe the preparation process of imaging ArabidopsisSAMs and primordia using a confocal microscope. We describe steps for dissection, visualization of meristems using dyes and fluorescent proteins, and gain 3D morphology of meristems. We then detail analysis of shoot meristems using time-lapse imaging.

How appendages, such as plant leaves or animal limbs, develop asymmetric shapes remains a fundamental question in biology. Although ongoing research has revealed the genetic regulation of organ pattern formation, how gene activity ultimately directs organ shape remains unclear. Here, we show that leaf dorsoventral (adaxial-abaxial) polarity signals lead to mechanical heterogeneity of the cell wall, related to the methyl-esterification of cell-wall pectins in tomato and Arabidopsis. Numerical simulations predicate that mechanical heterogeneity is sufficient to produce the asymmetry seen in planar leaves. Experimental tests that alter pectin methyl-esterification, and therefore cell wall mechanical properties, support this model and lead to polar changes in gene expression, suggesting the existence of a feedback mechanism for mechanical signals in morphogenesis. Thus, mechanical heterogeneity within tissue may underlie organ shape asymmetry. Mechanical heterogeneity due to differences in methyl-esterification of cell wall pectins is sufficient to produce asymmetric leaf morphology, as predicted by a mathematical model and validated in Arabidopsisand tomato.

For ℓ 0 penalized (nonlinear) regression problems, most existing algorithms carried out theoretical analysis and numerical calculation with a fixed step size. However, the selection of an appropriate step size and the guarantee of good performance depend heavily on the parameters of the restricted strong convexity and smoothness of the loss function, which are difficult to calculate in practice. To overcome this problem, a novel and efficient data-driven line search rule is proposed to adaptively determine the appropriate step size based on the idea of support detection and root finding. For the step size by the line search, the ℓ 2 error bound of iteration sequence and the target regression coefficient has be analyzed without any restrictions on the parameters of the loss function. A lot of numerical comparisons with state-of-the-art algorithms in linear and logistic regression problems show the stability, effectiveness and superiority of the proposed algorithms. [ABSTRACT FROM AUTHOR]

Abstract: Cell is the functional unit of life. To study the complex interactions of systems of biological molecules, it is crucial to dissect these molecules at the cell level. In recent years, major progresses have been made by plant biologists to profile gene expression in specific cell types at the genome-wide level. Approaches based on the isolation of cells, polysomes or nuclei have been developed and successfully used for studying the cell types from distinct organs of several plant species. These cell-level data sets revealed previously unrecognized cellular properties, such as cell-specific gene expression modules and hormone response centers, and should serve as essential resources for functional genomic analyses. Newly developed technologies are more affordable to many laboratories and should help to provide new insights at the cellular resolution in the near future.

We propose an enhanced support detection and root finding approach (ESDAR) to variable selection in sparse linear models. ESDAR is motivated from the KKT conditions for the ℓ 0 penalized regression. In ESDAR, we introduce a step size to balance the primal and dual variables in determining the support of the solution. We establish a sharp oracle error bound and an oracle support recovery property for the solution sequence generated by ESDAR under the weakest possible condition on the design matrix that is sufficient and necessary for the model to be identifiable. The conditions for the oracle results we obtained are weaker than those for Lasso and concave selection methods including SCAD and MCP in the literature. [ABSTRACT FROM AUTHOR]

Cortical microtubules (CMTs) play pivotal roles during plant cell growth and division. The organization of CMTs undergoes important changes during different cellular and developmental processes. Here, we describe two methods for the visualization of CMT organization in plant cells using confocal laser scanning microscopy. CMT networks in the outer tissue layers can be directly visualized by live imaging of a fluorescent reporter line, and a protocol combining sectioning and immunostaining is applied for visualization of CMTs throughout tissues.

Background Transposable element (TE)-related genes comprise a significant portion of the gene catalog of grasses, although their functions are insufficiently characterized. The recent availability of TE-related gene annotation from the complete genome sequence of rice (Oryza sativa) has created an opportunity to conduct a comprehensive evaluation of the transcriptional activities of these potentially mobile elements and their related genes.Results We conducted a genome-wide survey of the transcriptional activity of TE-related genes associated with 15 developmental stages and stress conditions. This dataset was obtained using a microarray encompassing 2,191 unique TE-related rice genes, which were represented by oligonucleotide probes that were free from cross-hybridization. We found that TE-related genes exhibit much lower transcriptional activities than do non-TE-related genes, although representative transcripts were detected from all superfamilies of both type I and II TE-related genes. The strongest transcriptional activities were detected in TE-related genes from among the MULE and CACTA superfamilies. Phylogenetic analyses suggest that domesticated TE-related genes tend to form clades with active transcription. In addition, chromatin-level regulations through histone and DNA modifications, as well as enrichment of certain cis elements in the promoters, appear to contribute to the transcriptional activation of representative TE-related genes.Conclusion Our findings reveal clear, albeit low, general transcription of TE-related genes. In combination with phylogenetic analysis, transcriptional analysis has the potential to lead to the identification of domesticated TEs with adapted host functions.

Jump regression analysis (JRA) provides a useful tool for estimating discontinuous functional relationships between a response and predictors. Most existing JRA methods consider the problems where there is only one or two predictors. It is unclear whether these methods can be directly extended to cases where there are multiple predictors. A jump additive model and a jump-preserving backfitting procedure are proposed. Jump additive models have the appeal that they make no restrictive parametric assumptions and allow possible discontinuities in the functional relationships, as with univariate JRA methods, but unlike them, jump additive models easily accommodate multiple predictors and the effects of individual predictors on the response can still be visually interpreted, regardless of the number of predictors. The proposed fitting procedure achieves the jump-preserving property by adaptively choosing, in each iteration of the backfitting algorithm, among two one-sided local linear estimates and a two-sided local linear estimate. Theoretical justifications and numerical studies show that it works well in applications. The procedure is also illustrated in analyzing a real data set. [ABSTRACT FROM AUTHOR]

We develop a constructive approach for ℓ 0 -penalized estimation in the sparse accelerated failure time (AFT) model with high-dimensional covariates. The proposed approach is based on Stute's weighted least squares criterion combined with ℓ 0 -penalization. This method is a computational algorithm that generates a sequence of solutions iteratively, based on active sets derived from primal and dual information and root finding according to the Karush-Kuhn-Tucker (KKT) conditions. We refer to the proposed method as AFT-SDAR (for support detection and root finding). An important aspect of our theoretical results is that we directly concern the sequence of solutions generated based on the AFT-SDAR algorithm. We prove that the estimation errors of the solution sequence decay exponentially to the optimal error bound with high probability, as long as the covariate matrix satisfies a mild regularity condition which is necessary and sufficient for model identification even in the setting of high-dimensional linear regression. An adaptive version of AFT-SDAR is also proposed, i.e., AFT-ASDAR, which determines the support size of the estimated coefficient in a data-driven fashion. Simulation studies demonstrate the superior performance of the proposed method over the lasso and MCP in terms of accuracy and speed. The application of the proposed method is also illustrated by analyzing a real data set. [ABSTRACT FROM AUTHOR]