Search for: All records

Hyperspectral sensors acquire spectral responses from objects with a large number of narrow spectral bands. The large volume of data may be costly in terms of storage and computational requirements. In addition, hyperspectral data are often information-wise redundant. Band selection intends to overcome these limitations by selecting a small subset of spectral bands that provide more information or better performance for particular tasks. However, existing band selection techniques do not directly maximize the task-specific performance, but rather utilize hand-crafted metrics as a proxy to the final goal of performance improvement. In this paper, we propose a deep learning (DL) architecture composed of a constrained measurement learning network for band selection, followed by a classification network. The proposed joint DL architecture is trained in a data-driven manner to optimize the classification loss along band selection. In this way, the proposed network directly learns to select bands that enhance the classification performance. Our evaluation results with Indian Pines (IP) and the University of Pavia (UP) datasets show that the proposed constrained measurement learning-based band selection approach provides higher classification accuracy compared to the state-of-the-art supervised band selection methods for the same number of bands selected. The proposed method shows 89.08% and 97.78% overall accuracy scores for IP and UP respectively, being 1.34% and 2.19% higher than the second-best method.

 

Vascular bundles transport water and photosynthate to all organs, and increased bundle number contributes to crop lodging resistance. However, the regulation of vascular bundle formation is poorly understood in the Arabidopsis stem.

We report a novel semi‐dominant mutant with high vascular activity,hva‐d, showing increased vascular bundle number and enhanced cambium proliferation in the stem. The activation of a C2H2 zinc finger transcription factor,AT5G27880/HVA, is responsible for thehva‐dphenotype. Genetic, biochemical, and fluorescent microscopic analyses were used to dissect the functions of HVA.

HVA functions as a repressor and interacts with TOPLESS via the conserved Ethylene‐responsive element binding factor‐associated Amphiphilic Repression motif. In contrast to the HVA activation line, knockout ofHVAfunction with a CRISPR‐Cas9 approach or expression of HVA fused with an activation domain VP16 (HVA‐VP16) resulted in fewer vascular bundles. Further, HVA directly regulates the expression of the auxin transport efflux facilitatorPIN1, as a result affecting auxin accumulation. Genetics analysis demonstrated that PIN1 is epistatic to HVA in controlling bundle number.

This research identifies HVA as a positive regulator of vascular initiation through negatively modulating auxin transport and sheds new light on the mechanism of bundle formation in the stem.

 

Fe^II‐ and α‐ketoglutarate‐dependent halogenases and oxygenases can catalyze site‐selective functionalization of C−H bonds via a variety of C−X bond forming reactions, but achieving high chemoselectivity for functionalization using non‐native functional groups remains rare. The current study shows that directed evolution can be used to engineer variants of the dioxygenase SadX that address this challenge. Site‐selective azidation of succinylated amino acids and a succinylated amine was achieved as a result of mutations throughout the SadX structure. The installed azide group was reduced to a primary amine, and the succinyl group required for azidation was enzymatically cleaved to provide the corresponding amine. These results provide a promising starting point for evolving additional SadX variants with activity on structurally distinct substrates and for enabling enzymatic C−H functionalization with other non‐native functional groups.

 

Fe^II‐ and α‐ketoglutarate‐dependent halogenases and oxygenases can catalyze site‐selective functionalization of C−H bonds via a variety of C−X bond forming reactions, but achieving high chemoselectivity for functionalization using non‐native functional groups remains rare. The current study shows that directed evolution can be used to engineer variants of the dioxygenase SadX that address this challenge. Site‐selective azidation of succinylated amino acids and a succinylated amine was achieved as a result of mutations throughout the SadX structure. The installed azide group was reduced to a primary amine, and the succinyl group required for azidation was enzymatically cleaved to provide the corresponding amine. These results provide a promising starting point for evolving additional SadX variants with activity on structurally distinct substrates and for enabling enzymatic C−H functionalization with other non‐native functional groups.

 

High-throughput next generation sequencing of cDNA, i.e. RNA-Seq, presents an unprecedented resource for characterizing the alternative splicing (AS) in complex eukaryotic transcriptomes. Accumulating evidences indicate that AS is developmentally regulated, but the precise responses of AS event to development is not well understood. Here, we describe a new method, based on an adjusted beta-distribution model, for detection of differential AS patterns from RNA-Seq data comparisons. Applying our method to two datasets of RNA-Seq for zika infection in human cells and pollen tissue in Arabidopsis thaliana, we identified 1,871 differentially AS events for 1,394 protein-coding genes in human and 496 differentially AS events for 358 protein-coding genes in Arabidopsis, respectively. The results included known AS events reported before as well as novel events, which demonstrate that the biological replicates are important in the effective identification using β-distribution. With a high accurate rate, our new method in differential AS identification will facilitate future investigation on transcriptomic annotation. 

Pre‐mRNA splicing is an essential step for the regulation of gene expression. In order to specifically capture splicing variants in plants for genome‐wide association studies (GWAS), we developed a software tool to quantify and visualise Variations of Splicing in Population (VaSP).

VaSP can quantify splicing variants from short‐read RNA‐seq datasets and discover genotype‐specific splicing (GSS) events, which can be used to prioritise causal pre‐mRNA splicing events in GWAS. We applied our method to an RNA‐seq dataset with 328 samples from 82 genotypes from a rice diversity panel exposed to optimal and saline growing conditions.

In total, 764 significant GSS events were identified in salt stress conditions. GSS events were used as markers for a GWAS with the shoot Na⁺accumulation, which identified six GSS events in five genes significantly associated with the shoot Na⁺content. Two of these genes,OsNUC1andOsRAD23emerged as top candidate genes with splice variants that exhibited significant divergence between the variants for shoot growth under salt stress conditions.

VaSP is a versatile tool for alternative splicing analysis in plants and a powerful tool for prioritising candidate causal pre‐mRNA splicing and corresponding genomic variations in GWAS.