Multiple Environments Research Articles

Robust and smart: Inference on phenotypic plasticity of Coffea canephora reveals adaptation to alternative environments

AbstractCoffee is an important crop with economic and social importance in several countries. With a daily consumption estimated at 2.2 billion cups, its sustainability is facing critical challenges given the projected climate changes. Coffea arabica, which represents ∼60% of the global market coffee is a delicate crop, quite susceptible to diseases and biotic stresses. Developing climate‐resilience cultivars is necessary, and it includes coffee plants adapted to new farming conditions that can meet the demand for biotic and abiotic tolerance and quality. In this context, Coffea canephora emerges as a potential candidate if the crop combines plasticity and cupping quality. Plant plasticity refers to adjusted phenotypic performance when grown in different environments, a fact that may help mitigate the detrimental effect of climate changes. In this study, using a multiple environment trial, we combined genomic and genotype‐by‐environment analyses to answer the following main question: How the climate effects may affect the phenotypic plasticity in C. canephora? Our contributions in this paper are fourfold: (i) we draw attention to the cupping quality and yield performance of C. canephora cultivars when evaluated in high‐altitude and cold weather, (ii) we compared C. arabica and C. canephora phenotypic plasticity and highlight genotypes with broad and specific adaptation to certain environmental conditions, and finally, (iii) using stochastic simulation, we emphasize the potential of molecular breeding in the long term in coffee. Altogether, we present an emerging view on how C. canephora could be a valid alternative for climate‐smart cultivars in a projected scenario of altered climatic conditions.

Identification, validation and candidate gene analysis of major QTL for Supernumerary spikelets in wheat

BackgroundThe number of spikelets per spike is a key trait that affects the yield of bread wheat (Triticum aestivum L.). Identification of the QTL for spikelets per spike and its genetic effects that could be used in molecular assistant breeding in the future.ResultsIn this study, four recombinant inbred line (RIL) populations were generated and used, having YuPi branching wheat (YP), with Supernumerary Spikelets (SS) phenotype, as a common parent. QTL (QSS.sicau-2 A and QSS.sicau-2D) related to SS trait were mapped on chromosomes 2 A and 2D through bulked segregant exome sequencing (BSE-Seq). Fourteen molecular markers were further developed within the localization interval, and QSS.sicau-2 A was narrowed to 3.0 cM covering 7.6 Mb physical region of the reference genome, explaining 13.7 − 15.9% the phenotypic variance. Similarly, the QSS.sicau-2D was narrowed to 1.8 cM covering 2.4 Mb physical region of the reference genome, and it explained 27.4 − 32.9% the phenotypic variance. These two QTL were validated in three different genetic backgrounds using the linked markers. QSS.sicau-2 A was identified as WFZP-A, and QSS.sicau-2D was identified a novel locus, different to the previously identified WFZP-D. Based on the gene expression patterns, gene annotation and sequence analysis, TraesCS2D03G0260700 was predicted to be a potential candidate gene for QSS.sicau-2D.ConclusionTwo significant QTL for SS, namely QSS.sicau-2 A and QSS.sicau-2D were identified in multiple environments were identified and their effect in diverse genetic populations was assessed. QSS.sicau-2D is a novel QTL associated with the SS trait, with TraesCS2D03G0260700 predicted as its candidate gene.

A Multiple Environment Available Path Planning Based on an Improved A* Algorithm

The objective of the path planning for a mobile robot is to generate a collision-free path from a starting position to a target position, aiming to realize a higher quality of path planning, an improved A* algorithm and a hybrid approach incorporating the dynamic window algorithm have been proposed for robot path planning in various environments in this paper. In global path planning, first, a bidirectional search strategy was introduced into to improve the searching efficiency, and an adaptive heuristic function was designed to reduce redundant search nodes. In the meantime, a filtering function for key path nodes and an enhanced jump point optimization method help to remove redundant nodes in the path, reduce turning angles, greatly shorten the path length, and smooth the path using cubic B-spline curves. Furthermore, in local path planning, the combination of key path nodes and the dynamic window approach (DWA) algorithm is utilized to achieve obstacle avoidance in dynamic environments and adjust the heading angle of the section enables seamless locomotion of the robot. Finally, the simulation experiments and physical experiments on the robot were conducted to validate that the proposed improved algorithm significantly improves the speed of path planning while also reducing the length of the planned path and improve the reliability of the algorithm when compared with other algorithms.

Racial Composition of Past and Current Social Environments and Health Literacy.

Research is needed to understand the impact of social determinants of health on health literacy throughout the life course. This study examined how racial composition of multiple past and current social environments was related to adults' health literacy. In this study, 546 adult patients at a primary care clinic in St. Louis, Missouri, completed a self-administered written questionnaire that assessed demographic characteristics and a verbally administered component that assessed health literacy with the Rapid Estimate of Adult Literacy in Medicine - Revised (REALM-R) and Newest Vital Sign (NVS), and self-reported racial composition of six past and four current social environments. Multilevel logistic regression models were built to examine the relationships between racial composition of past and current social environments and health literacy. Most participants identified as Black or multiracial (61%), had a high school diploma or less (54%), and household income <$20,000 (72%). About 56% had adequate health literacy based on REALM-R and 38% based on NVS. In regression models, participants with multiple past white environments (e.g., locations/conditions in which most of the people who live, go to school, work, and have leisure time are White) and (vs. 0 or 1) were more likely to have adequate health literacy based on REALM-R (adjusted odds ratio [aOR] = 1.79; 95% confidence interval [CI]: 1.04-3.07). Similarly, participants who had multiple past white social environments were more likely (aOR = 1.94, 95% CI: 1.15-3.27) to have adequate health literacy based on NVS than those who had not. The racial composition of current social environments was not significantly associated with health literacy in either model. Racial composition of past, but not current, educational and residential social environments was significantly associated with adult health literacy. The results highlight the importance of examining the impact of social determinants over the life course on health literacy. The findings suggest that policies ensuring equitable access to educational resources in school and community contexts is critical to improving equitable health literacy. [HLRP: Health Literacy Research and Practice. 2024;8(3):e130-e139.].

Brain-Inspired Deep Meta-Reinforcement Learning for Active Coordinated Fault-Tolerant Load Frequency Control of Multi-Area Grids

This paper proposes an active coordinated fault tolerance load frequency control (AFCT-LFC) method, which effectively prevents sudden frequency changes caused by unit actuator failures or unplanned decommissioning in a multi-area interconnected grid subject to the performance-based frequency regulation market mechanism. It can also reduce regulation mileage payments and achieve multi-objective active fault-tolerant control. In addition, this paper proposes a brain-Inspired deep meta-deterministic policy gradient algorithm (BIMA-DMDPG), which adopts multi-agent centralized training, equates the controller of each area as an agent capable of independent decision making, and implements distributed training by dividing the environment into multiple environments. In addition, meta-reinforcement learning is employed to realize multi-task collaborative learning. The optimal policy is actively selected under different fault conditions to achieve active fault-tolerant control. The superior performance of the method is verified in a four-area LFC model of the China Southern Grid (CSG), in which it is tested alongside a selection of existing algorithms. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Note to Practitioners</i> —AFCT-LFC is based on advanced artificial intelligence algorithm, which can effectively identify any fault in multi-area grids and make rapid response to achieve active fault-tolerant control. Compared with the existing model-based fault-tolerant control methods, the BIMA-DMDPG algorithm proposed in this paper does not need to rely on accurate mathematical models, and can be applied to practice through simple training, which is very suitable for practical applications. Therefore, AFCT-LFC is an advanced adaptive active fault-tolerant control method that can be truly applied in practice because of its fast-decision-making ability and performance.

Shrinkage separation prediction of concrete-filled steel tube arch bridge: A coupling model concerning multiple factors

The shrinkage separation is one of the key factors restricting the development of concrete-filled steel tube (CFST) arch bridge. In this paper, its mechanism and evaluation method concerning multiple factors, including structure, environment, material and construction were investigated. Firstly, the hydration process of in-tube concrete was simulated using the hydration degree as a basic state parameter. Then with the consideration of energy and mass conservation, as well as the moisture and heat transmission effects, the temperature and relative humidity fields, early-age mechanical properties, basic creep, and total deformation of in-tube concrete were further calculated. On the basis, in accordance with the stress criterion and displacement coordination, a multi-field (hydro-thermo-hygro-constraint) coupling model was finally established, taking the ratio of normal stress to bonding strength at the steel-concrete interface as the shrinkage separation risk index. The simulation results obtained from the model were in good agreements with the full-scale CFST model test.

Influence of Grease Contamination on Acoustic Emission Monitoring of Low-speed Roller Bearings

Large-scale low-speed rolling element bearings form crucial connections in offshore installations such as heavy-lifting cranes, single point mooring systems, and wind turbines. Due to the stochastic nature of wind and waves, the applied loads on these bearings are hardly predictable. Furthermore, the remote nature of the offshore environment requires a high standard of operational safety and reliability, reinforcing the need for advanced inspection and monitoring methods. Acoustic emission (AE) is a continuous monitoring technique for condition assessment of low-speed bearings, as it may detect the stress waves associated with developing degradation within the rolling elements. This paper presents an investigation on the influence of grease contamination on acoustic emission (AE) monitoring of low-speed bearings. It discusses several experiments that have been conducted in various regimes of particle contaminated lubrication, and proposes a strategy for condition monitoring. The presented experiments have been conducted with differing bearing geometries in multiple testing environments. The contamination particles have been both naturally developed and artificially introduced. The results total of 9 experimental cases are discussed and compared. All experiments follow the same data-acquisition principles, utilising arrays of three AE transducers sensitive in a range between 40-580 kHz. Data is recorded while operating the bearing at low speed under load. Abrasive wear of the rolling elements and crushing of the particles are concluded to be the main sources of AE due to particle contamination. Processing consists of waveform cross-correlation clustering and feature analysis. The results suggest that hard abrasive particles make significant contribution to the AE signals, which is associated with abrasive wear of the rollers and raceways. Comparisons with the contribution of softer and finer particles are presented as well.

Team resilience in multiple project environments: What characteristics should be explored?

Purpose: This paper aims to explore how project team members' resilience differs based on industry type, project budget, company type, and gender. Design/methodology/approach: Using the conservation of resources theory, this study focuses on exploring the insight of the gathered data from the perspective of descriptive statistics. To achieve this objective, a descriptive analysis was conducted using purposive sampling and snowball techniques, gathering data from 349 respondents. The collected data was analyzed using IBM SPSS software version 25. Findings: The study's results indicate that male respondents are dominant, primarily working in the construction industry, while female respondents tend to work in other industry sectors. Moreover, male respondents exhibited higher resilience scores, but no significant difference in resilience was observed based on gender. However, resil-ience levels varied according to the industry sector. Respondents from industries other than the construction sector tended to display higher resilience. Research limitations/implications: Considering that the study finds only minor variations, businesses in the same industrial sector might think about implementing a practice benchmark. Examining businesses outside of these im-portant industries, however, may provide difficulties. Subsequent data-gathering endeavors may be broadened to encompass more industry sectors and subsequently scrutinized via case studies utilizing smaller samples from certain sectors. Originality/value: This study reveals that it is important to further expand and investigate these insights. Resources for team resilience, such as team confidence, a roadmap for teamwork, the ability to improvise as a team, and psychological safety for the team, can be added to future research. Furthermore, it is crucial to do quantitative research to look at how these elements show up in various industry sectors.

STFGCN: Spatial–temporal fusion graph convolutional network for traffic prediction

Accurate traffic prediction plays a crucial role in improving traffic conditions and optimizing road utilization. Effectively capturing the multi-scale temporal dependencies and dynamic spatial dependencies is crucial for accurate traffic prediction. These features can effectively reflect complex dynamic spatial–temporal processes, which have not been comprehensively addressed in most existing research work. Motivated by this issue, the primary contribution of this paper lies in proposing a novel Spatial–Temporal Fusion Graph Neural Network (STFGCN) for accurate traffic prediction, achieved by extracting multi-scale temporal dependencies from multiple semantic environments and constructing a dynamic adaptive graph to model spatial dependencies based on temporal characteristics. Specifically, to capture the multi-scale dynamic temporal dependencies effectively, a Multi-Scale Fusion Convolution (MSFC) module is designed, in which the temporal dependencies are extracted from multiple textual environments by utilizing multi-scale convolution. In order to model dynamic spatial dependencies, a Spatial Adaptive Fusion Convolution (SAFC) module is designed by combining the recent coherence and periodicity to infer dynamic graphs, which are then fused to model dynamic spatial dependencies. Extensive experimental results on five real-world datasets demonstrate that the proposed STFGCN has superior performance. Specifically, compared with the state-of-the-art baselines, STFGCN reduced 1.2% to 16.4% in RMSE measure.

Queering child welfare and juvenile justice: Recommendations for affirming LGBTQ+ youth

AbstractLesbian, gay, bisexual, transgender, and queer (LGBTQ+) people have historically been subjected to maltreatment across multiple environments and systems – including their homes, schools, neighborhoods, workplaces, healthcare facilities, and more. The criminal justice and child welfare systems were built on the foundations of cisnormativity and heteronormativity, making LGBTQ+ youth susceptible to systemic oppression and discrimination when compared to their youth counterparts. As such, this paper examines the unique experiences of LGBTQ+ youth in the juvenile and child welfare systems by discussing the systemic inequities and unique challenges that exist for this population – beginning with the risk factors that make them susceptible to exposure to both systems in general. Furthermore, we provide practical recommendations for legal and clinical professionals to better serve their LGBTQ+ youth clients – ranging from changes in policies and legislation, increased staff training, and strategies to create safe and inclusive environments for all youth.

Combining genome-wide association study and linkage mapping in the genetic dissection of amylose content in maize (Zea mays L.).

Integrated linkage and association analysis revealed genetic basis across multiple environments. The genes Zm00001d003102 and Zm00001d015905 were further verified to influence amylose content using gene-based association study. Maize kernel amylose is an important source of human food and industrial raw material. However, the genetic basis underlying maize amylose content is still obscure. Herein, we used an intermated B73 × Mo17 (IBM) Syn10 doubled haploid population composed of 222 lines and a germplasm set including 305 inbred lines to uncover the genetic control for amylose content under four environments. Linkage mapping detected 16 unique QTL, among which four were individually repeatedly identified across multiple environments. Genome-wide association study revealed 17 significant (P = 2.24E-06) single-nucleotide polymorphisms, of which two (SYN19568 and PZE-105090500) were located in the intervals of the mapped QTL (qAC2 and qAC5-3), respectively. According to the two population co-localized loci, 20 genes were confirmed as the candidate genes for amylose content. Gene-based association analysis indicated that the variants in Zm00001d003102 (Beta-16-galactosyltransferase GALT29A) and Zm00001d015905 (Sugar transporter 4a) affected amylose content across multi-environment. Tissue expression analysis showed that the two genes were specifically highly expressed in the ear and stem, respectively, suggesting that they might participate in sugar transport from source to sink organs. Our study provides valuable genetic information for breeding maize varieties with high amylose.

Multi-GWAS reveals significant genomic regions for Mungbean yellow mosaic India virus resistance in urdbean (Vigna mungo (L.) across multiple environments.

Unraveling genetic markers for MYMIV resistance in urdbean, with 8 high-confidence marker-trait associations identified across diverse environments, provides crucial insights for combating MYMIV disease, informing future breeding strategies. Globally, yellow mosaic disease (YMD) causes significant yield losses, reaching up to 100% in favorable environments within major urdbean cultivating regions. The introgression of genomic regions conferring resistance into urdbean cultivars is crucial for combating YMD, including resistance against mungbean yellow mosaic India virus (MYMIV). To uncover the genetic basis of MYMIV resistance, we conducted a genome-wide association study (GWAS) using three multi-locus models in 100 diverse urdbean genotypes cultivated across six individual and two combined environments. Leveraging 4538 high-quality single nucleotide polymorphism (SNP) markers, we identified 28 unique significant marker-trait associations (MTAs) for MYMIV resistance, with 8 MTAs considered of high confidence due to detection across multiple GWAS models and/or environments. Notably, 4 out of 28 MTAs were found in proximity to previously reported genomic regions associated with MYMIV resistance in urdbean and mungbean, strengthening our findings and indicating consistent genomic regions for MYMIV resistance. Among the eight highly significant MTAs, one localized on chromosome 6 adjacent to previously identified quantitative trait loci for MYMIV resistance, while the remaining seven were novel. These MTAs contain several genes implicated in disease resistance, including four common ones consistently found across all eight MTAs: receptor-like serine-threonine kinases, E3 ubiquitin-protein ligase, pentatricopeptide repeat, and ankyrin repeats. Previous studies have linked these genes to defense against viral infections across different crops, suggesting their potential for further basic research involving cloning and utilization in breeding programs. This study represents the first GWAS investigation aimed at identifying resistance against MYMIV in urdbean germplasm.

Development of an automated extraction and radiocarbon dating method for fossil pollen deposited in lake Motosu, Japan

Recently, radiocarbon (14C) dating methods using fossil pollen extracted from sediments with a flow cytometer (cell sorter) are under development. Technical limitations experienced by previous studies required extraction of grains <80 μm in diameter. Thus, obtaining a sufficient mass of carbon for dating requires extracting a very large number (∼105) of grains. Another challenge faced by earlier work was preventing contamination by exotic carbon during the extraction process. Here we present a novel solution to this problem by using a cell sorter with a newly designed pretreatment method and an improved extraction method. This enables us to extract large pollen fossils than was previously possible. By using grains, >100 μm in diameter, such as Pinus sp., we have reduced the number of grains for required for 14C dating by an order of magnitude, particularly when considering the recent advances in measure ultra-small carbon masses on a single-stage accelerator mass spectrometer at the Atmosphere and Ocean Research Institute, University of Tokyo. We then apply this method to sediments recovered from Lake Motosu, which already has a very robust chronology, to evaluate the new method. Results indicate the method is successful and reveal temporal radiocarbon reservoir effects that appear related to the changes in the depositional environment and/or hydroclimate. The method presented here is widely applicable across multiple environments.

Deconstructing the physical output between tapering sessions, match days, and post-match training sessions: A single case report

Talented adolescent professional soccer players can experience multiple within-career transitions during the course of a competitive season in Major League Soccer. Consequently, players may shift between multiple team environments, including the first team (FT), reserve team (RT) and youth national team (NT). When changing environments, players must adapt and cope with increased physical demands associated with training and competing in the FT (Morris et al., 2016, J Appl Sport Psych, 28, 375-391), as well as the NT (McKay et al., 2022, J Appl Sport Psych, 34, 1272-1294). Consequently, the aim of this single case report was to analyse the physical output during matches, tapering and post-match sessions from a professional adolescent shifting between multiple team environments. External load data were collected via Global Positioning Systems (GPS) for an entire competitive season in three different environments (FT, RT, NT) across tapering sessions (MD-2 and MD-1), match days (MD), and post-match training sessions (MD+1/+2; ≥60 and &lt;60 minutes played). The following intensity related measures (i.e., relative distance covered in meters, which is divided by the duration of the training session or match in minutes) were extracted for analysis: total distance covered, distance covered in multiple speed zones (Zone 1: 0-0.6 km/h; Zone 2: 0.7-7.1 km/h; Zone 3: 7.2-14.3 km/h; Zone 4: 14.4-.19.7 km/h; Zone 5: 19.8-25.1 km/h; Zone 6: &gt;25.1 km/h), moderate speed running distance (&gt;Z4; MSRD), and high speed running distance (&gt;Z5; HSRD). A multivariate analysis of variance with physical output variables as dependent variables and session type (MD+/-) as fixed factors was conducted. The results indicated that the session type had a significant effect on all variables (F8,128 = 22.14-158.86; P &lt; 0.01; ES = 0.41-0.83), as well as the environment (F2,128 = 3.8-26.7; P &lt; 0.5; ES = 0.06-0.30) except relative distance covered in Z6 and SPD (P &gt; 0.05). However, the interaction between session type and environment had no significant effect on any variable (P &gt; 0.05), except for total distance (F8,128 = 2.33; P = 0.02; ES = 0.13) and distance covered in Z1 (F8,128 = 5.32; p&lt;0.001; ES = 0.25). In conclusion, the adolescent professional player was exposed to different relative intensities in training and matches across multiple team environments over the course of a competitive season. In particular, the adolescent player recorded greater relative intensities in the FT environment relative to the RT and NT environments.

Multi-fidelity reinforcement learning with control variates

In many computational science and engineering applications, the output of a system of interest corresponding to a given input can be queried at different levels of fidelity with different costs. Typically, low-fidelity data is cheap and abundant, while high-fidelity data is expensive and scarce. In this work we study the reinforcement learning (RL) problem in the presence of multiple environments with different levels of fidelity for a given control task. We focus on improving the RL agent’s performance with multifidelity data. Specifically, a multifidelity estimator that exploits the cross-correlations between the low- and high-fidelity returns is proposed to reduce the variance in the estimation of the state–action value function. The proposed estimator, which is based on the method of control variates, is used to design a multifidelity Monte Carlo RL (MFMCRL) algorithm that improves the learning of the agent in the high-fidelity environment. The impacts of variance reduction on policy evaluation and policy improvement are theoretically analyzed by using probability bounds. Our theoretical analysis and numerical experiments demonstrate that for a finite budget of high-fidelity data samples, our proposed MFMCRL agent attains superior performance compared with that of a standard RL agent that uses only the high-fidelity environment data for learning the optimal policy. 22This article is an extended version of a paper originally presented at the ESANN (European Symposium on Artificial Neural Networks) conference in 2023.

Identification of a pleiotropic QTL and development KASP markers for HPW, HSW, and SP in peanut

High yield has always been the primary objective of peanut breeding. 100-pod weight (HPW), 100-seed weight (HSW), and shelling percentage (SP), are crucial components of peanut yield. This study aimed to construct a high-density linkage map by resequencing the recombined inbred lines (RILs) derived from a cross between “Silihong” (A. hypogaea var. fastigiate) and “Jinonghei 3” (A. hypogaea var. hypogaea). The map consisted of 4,499 bins spread across 20 chromosomes, totaling 1712.32 cM in length with an average inter-marker distance of 0.38 cM. A total of 46 quantitative trait loci (QTLs) were identified across three environments. The major QTLs, including qHPW5.2, qHPW18.1, qSP7.1, qSP8.1, qSP8.2, qSP18.1, and qSP18.2, exhibited PVE (phenotypic variation explained) of 12.04, 11.41, 16.53, 24.17, 10.49, 10.82, and 29.89%, respectively. Fourteen QTLs identified across multiple environments were considered stable. One QTL (qHPW7, qHSW7.1, qSP7) was associated with all three traits, with the PVE value of 8.91, 9.04, and 16.53% for HPW, HSW, and SP, respectively. The genome-wide association study was conducted using the US mini-core collection to validate the accuracy of QTL mapping. Across two environments, 115 single-nucleotide polymorphisms (SNPs) were significantly associated with HPW, HSW, and SP in the association panel. Six SNPs were associated with two traits, explaining an average phenotypic variation of 13.84%. Combining the two mapping populations, AX-176802178, detected on chromosome 7 in the association panel, which controlled SP, was located within the QTL qSP7 confidence interval defined by the RILs. Moreover, three KASP markers were developed and validated in peanut landraces and varieties. These QTLs might offer valuable insights for understanding the genetic basis of HPW, HSW, and SP and provide useful molecular markers for marker-assisted breeding in peanuts.

Reinforcement learning for closed-loop regulation of cardiovascular system with vagus nerve stimulation: a computational study

Objective. Vagus nerve stimulation (VNS) is being investigated as a potential therapy for cardiovascular diseases including heart failure, cardiac arrhythmia, and hypertension. The lack of a systematic approach for controlling and tuning the VNS parameters poses a significant challenge. Closed-loop VNS strategies combined with artificial intelligence (AI) approaches offer a framework for systematically learning and adapting the optimal stimulation parameters. In this study, we presented an interactive AI framework using reinforcement learning (RL) for automated data-driven design of closed-loop VNS control systems in a computational study. Approach. Multiple simulation environments with a standard application programming interface were developed to facilitate the design and evaluation of the automated data-driven closed-loop VNS control systems. These environments simulate the hemodynamic response to multi-location VNS using biophysics-based computational models of healthy and hypertensive rat cardiovascular systems in resting and exercise states. We designed and implemented the RL-based closed-loop VNS control frameworks in the context of controlling the heart rate and the mean arterial pressure for a set point tracking task. Our experimental design included two approaches; a general policy using deep RL algorithms and a sample-efficient adaptive policy using probabilistic inference for learning and control. Main results. Our simulation results demonstrated the capabilities of the closed-loop RL-based approaches to learn optimal VNS control policies and to adapt to variations in the target set points and the underlying dynamics of the cardiovascular system. Our findings highlighted the trade-off between sample-efficiency and generalizability, providing insights for proper algorithm selection. Finally, we demonstrated that transfer learning improves the sample efficiency of deep RL algorithms allowing the development of more efficient and personalized closed-loop VNS systems. Significance. We demonstrated the capability of RL-based closed-loop VNS systems. Our approach provided a systematic adaptable framework for learning control strategies without requiring prior knowledge about the underlying dynamics.

Comprehensive diallel analysis of combining ability and heterosis to Fusarium verticillioides ear rot resistance in maize

AbstractMaize (Zea mays L.) Fusarium ear rot (FER), which is caused by the fungal pathogen Fusarium verticillioides, is a major ear disease affecting maize production in China. Developing resistant maize hybrids is a cost‐effective and environmentally friendly method of reducing yield losses due to FER. Using the nail punch inoculation method, 17 highly resistant and 16 highly susceptible inbred lines were identified in this study, and a large‐scale maize hybrid consisting of 528 F1’s was produced using a half‐diallel cross model. The hybrids were inoculated with F. verticillioides in 6 year–location environment combinations during 2020–2022. Our results showed that estimates of genotypic, environmental, and genotype × environment interaction variance components were very significant (p &lt; 0.001). All the crosses generated from highly resistant inbred lines exhibited high kernel resistance, whereas those generated from highly susceptible inbred lines mostly exhibited high susceptibility to FER. Crosses between resistant and susceptible inbred lines showed intermediate resistance to FER. The general combining ability (GCA) effect was greater than the specific combining ability effect across multiple environments, indicating that the inheritance of FER resistance was mainly influenced by additive genetic effects. The correlation coefficient between the expected GCA value and hybrid resistance was 0.789, whereas that between the mid‐parent value and hybrid resistance was only 0.644. Inbred performance per se and their corresponding hybrids across multiple environments were significantly correlated (r = 0.823). However, no relationship was observed between the FER resistance of F1’s and that of their parental heterotic groups.

Genome-Wide Association Study and Candidate Gene Mining of Seed Size Traits in Soybean

Seed size traits, including seed length (SL), seed width (SW), and seed thickness (ST), are crucial appearance parameters that determine soybean seed weight, yield, and ultimate utilization. However, there is still a large gap in the understanding of the genetic mechanism of these traits. Here, 281 soybeans were utilized to analyze the genetic architecture of seed size traits in different years through multiple (single-locus and multi-locus) genome-wide association study (GWAS) models, and candidate genes were predicted by integrating information on gene function and transcriptome sequencing data. As a result, two, seven, and three stable quantitative trait nucleotides (QTNs) controlling SL, SW, and ST were detected in multiple environments using the single-locus GWAS model, and concurrently detected by the results of the multi-locus GWAS models. These stable QTNs are located on 10 linkage disequilibrium blocks, with single genome regions ranging in size from 20 to 440 kb, and can serve as the major loci controlling soybean seed size. Furthermore, by combining gene functional annotation and transcriptome sequencing data of seeds at different developmental stages from two extreme soybean accessions, nine candidate genes, including Glyma.05G038000, Glyma.05G244100, Glyma.05G246900, Glyma.07G070200, Glyma.11G010000, Glyma.11G012400, Glyma.17G165500, Glyma.17G166500, and Glyma.20G012600 within the major loci that may regulate soybean seed size, were mined. Overall, these findings offer valuable insights for molecular improvement breeding as well as gene functional studies to unravel the mechanism of soybean seed size.

Local Structural Features Elucidate Crystallization of Complex Structures.

Complex crystal structures are composed of multiple local environments, and how this type of order emerges spontaneously during crystal growth has yet to be fully understood. We study crystal growth across various structures and along different crystallization pathways, using self-assembly simulations of identical particles that interact via multiwell isotropic pair potentials. We apply an unsupervised machine learning method to features from bond-orientational order metrics to identify different local motifs present during a given structure's crystallization process. In this manner, we distinguish different crystallographic sites in highly complex structures. Tailoring this order parameter to structures of varying complexity and coordination number, we study the emergence of local order along a multistep crystal growth pathway─from a low-density fluid to a high-density, supercooled amorphous liquid droplet and to a bulk crystal. We find a consistent under-coordination of the liquid relative to the average coordination number in the bulk crystal. We use our order parameter to analyze the geometrically frustrated growth of a Frank-Kasper phase and discover how structural defects compete with the formation of crystallographic sites that are more high-coordinated than the liquid environments. The method presented here for classifying order on a particle-by-particle level has broad applicability to future studies of structural self-assembly and crystal growth, and they can aid in the design of building blocks and for targeting pathways of formation of soft-matter structures.