Short-term Selection Research Articles

The genome-wide signature of short-term temporal selection

Despite evolutionary biology's obsession with natural selection, few studies have evaluated multigenerational series of patterns of selection on a genome-wide scale in natural populations. Here, we report on a 10-y population-genomic survey of the microcrustacean Daphnia pulex. The genome sequences of [Formula: see text]800 isolates provide insights into patterns of selection that cannot be obtained from long-term molecular-evolution studies, including the following: the pervasiveness of near quasi-neutrality across the genome (mean net selection coefficients near zero, but with significant temporal variance about the mean, and little evidence of positive covariance of selection across time intervals); the preponderance of weak positive selection operating on minor alleles; and a genome-wide distribution of numerous small linkage islands of observable selection influencing levels of nucleotide diversity. These results suggest that interannual fluctuating selection is a major determinant of standing levels of variation in natural populations, challenge the conventional paradigm for interpreting patterns of nucleotide diversity and divergence, and motivate the need for the further development of theoretical expressions for the interpretation of population-genomic data.

Client selection for federated learning using combinatorial multi-armed bandit under long-term energy constraint

In a federated learning system, it is often the case that the more clients it involves, the less increment of the outcome it achieves. It is thus essential to design a client selection strategy to choose an appropriate subset of the clients to participate in federated learning. However, client selection is not easy due to the heterogeneity of clients and the long-term energy budget of each client. Moreover, long-term energy budgets intertwined with the short-term client selection often make the problem NP-hard. In this paper, we propose an online strategy Energy-Aware Client Selection for Federated Learning (EACS-FL) to address this problem. The problem is formulated with a joint energy and delay optimization objective, and the Combinatorial Multi-Armed Bandit (CMAB) is introduced to solve the problem in an online manner. We take advantage of Lyapunov optimization to manage energy consumption of clients, which enables us to deal with independent energy budgets through minimizing virtual energy deficit queues. Theoretical analysis shows that EACS-FL achieves sublinear regret and keeps all queues stable. Experiment results exhibit that the proposed approach outperforms the existing works and achieves close-to-optimal delay and energy consumption performance.

Multi-objective reinforcement learning based on nonlinear scalarization and long-short-term optimization

PurposeMany practical control problems require achieving multiple objectives, and these objectives often conflict with each other. The existing multi-objective evolutionary reinforcement learning algorithms cannot achieve good search results when solving such problems. It is necessary to design a new multi-objective evolutionary reinforcement learning algorithm with a stronger searchability.Design/methodology/approachThe multi-objective reinforcement learning algorithm proposed in this paper is based on the evolutionary computation framework. In each generation, this study uses the long-short-term selection method to select parent policies. The long-term selection is based on the improvement of policy along the predefined optimization direction in the previous generation. The short-term selection uses a prediction model to predict the optimization direction that may have the greatest improvement on overall population performance. In the evolutionary stage, the penalty-based nonlinear scalarization method is used to scalarize the multi-dimensional advantage functions, and the nonlinear multi-objective policy gradient is designed to optimize the parent policies along the predefined directions.FindingsThe penalty-based nonlinear scalarization method can force policies to improve along the predefined optimization directions. The long-short-term optimization method can alleviate the exploration-exploitation problem, enabling the algorithm to explore unknown regions while ensuring that potential policies are fully optimized. The combination of these designs can effectively improve the performance of the final population.Originality/valueA multi-objective evolutionary reinforcement learning algorithm with stronger searchability has been proposed. This algorithm can find a Pareto policy set with better convergence, diversity and density.

Optimal open nucleus breeding system for long-term genetic gain in the Pacific white shrimp using genomic selection

The open nucleus breeding system (ONBS) has been widely employed in livestock breeding and has demonstrated a desirable genetic gain compared to the traditional closed nucleus breeding system (CNBS). However, no related study and application for ONBS exist on aquatic animal breeding. The present investigation simulates the family-based ONBS with different proportions of introduced multiplier individuals used as parents of the nucleus population (NP) and CNBS schemes for the harvest body weight of the Pacific white shrimp (Litopenaeus vannamei). The multiplier population (MP) comprised progenies from eight families with high estimated breeding value (EBV) or genomic EBV (GEBV) of NP of the previous generation. ONBS was assigned three proportions (10%, 30%, and 50%) of the introduced MP individuals. The short-term selection (G0 to G10) and long-term selection (G0 to G20) were performed for CNBS and ONBS schemes using pedigree-based best linear unbiased prediction (PBLUP) with the A relationship matrix from the pedigree or single-step genomic BLUP (ssGBLUP) with the H relationship matrix integrating SNP markers and pedigree. The genetic gain of body weight and inbreeding coefficient of NP for eight breeding schemes with CNBS_PBLUP (CP), CNBS_ssGLUP (CS), ONBS_PBLUP (OP10, OP30, OP50), and ONBS_ssGBLUP (OS10, OS30, OS50) with three proportions of the introduced MP individuals were calculated. Compared to CP, genetic gains for the three OP schemes increased by 7.68 to 11.27% after short-term selection, and by 2.51 to 5.79% after long-term selection. In comparison to CS, genetic gains for the three OS schemes increased by 8.21 to 8.96% after short-term selection, and from 4.46 to 5.87% after long-term selection. The ONBS schemes achieved higher genetic gains than the corresponding CNBS schemes whether using genomic or pedigree information. The advantage of ONBS over CNBS is reduced after long-term selection. Genetic gains for the three OS schemes increased by 9.77 to 14.06% after long-term selection compared with the corresponding OP schemes. Furthermore, no significant difference (P > 0.05) in genetic gain was revealed among the ONBS schemes after either pedigree-based or genomic selection. Inbreeding coefficients for the three OP schemes increased from 59.45 to 70.98% compared with CP, whereas inbreeding coefficients for the three OS schemes increased from 24.72 to 32.22% compared with CS. Besides, after long-term selection, inbreeding coefficients for the three OS schemes decreased from 13.13 to 14.79% compared with the corresponding OP schemes. In a nutshell, ONBS using introduced MP individuals has great promise in the selective breeding program of L. vannamei. Genomic information has more potential to improve selection response and control inbreeding level than pedigree information when applying ONBS.

Optimization of short-term stock selection based on volume and price using a non-cooperative parallel DEA model

This paper presents a novel approach to portfolio optimization in the field of finance, with a specific focus on short-term yield. Existing literature has mainly utilized fundamental data to predict long-term trends in stock prices, but our proposed methodology utilizes technical indicators based on the theory of chasing up. Furthermore, we address the non-cooperative nature of volume and price fluctuation indicators and introduce non-cooperative theory into the short-term volume and price stock selection scheme for the first time. We propose an optimization of short-term stock selection based on volume and price using a non-cooperative parallel Data Envelopment Analysis (DEA) model, which we apply to Chinese main board listed companies. Our empirical results demonstrate the effectiveness of our model in selecting high-yield stocks in the short term. This paper contributes to the ongoing discussion on portfolio optimization and presents a compelling solution for investors seeking to maximize their financial gains. The proposed methodology can be utilized in practical applications and has significant implications for the financial industry.

Short term selection response on growth and reproductive traits in broiler lines of chicken

Broiler chicken is highly demanded to meet proteins needs of growing population. This research was designed to evaluate the effect of selection response on the growth and reproductive performance traits of broiler chicken lines. Broiler chickens were selected into male and female lines. Body weight, conformation traits, production and reproductive traits were evaluated. Heritability and selection differential were computed and were used to estimate selection response. Correlations among serum alkaline phosphatase (SAP), total protein (TP) and Acetylcholine esterase (AchE) and production and reproductive traits were estimated. Higher heritability estimates were obtained in the body weight and thigh length of the selected lines. Large selection differentials were obtained for body weight and body conformation traits. Responses to selection were largest in the selected lines, Highest and significant number of egg was reported in the dam selected line. The selected lines revealed higher egg weight. SAP was positively and significantly correlated with egg weight (r = 0.761, p = 0.001), fertility (r = 0.870, p = 0.0001) and hatchability (r = 0.915, p = 0.0001). Total protein and Acetylcholine esterase were positively correlated with egg number. It is therefore concluded in this study that, selection, could lead to improvement in growth and reproduction traits in broiler chicken and SAP, TP and AchE could be used as marker for selecting egg weight and egg number in broiler chicken. Le poulet de chair est très demandé pour répondre aux besoins en protéines d'u caractéristiques de croissance et de performance reproductive des lignées de poulets de chair. Les poulets de chair ont été sélectionnés en lignées mâles et femelles. Le poids corporel, les traits de conformation, les traits de production et de reproduction ont été évalués. L'héritabilité et le différentiel de sélection ont été calculés et utilisés pour estimer la réponse de sélection. Les corrélations entre la phosphatase alcaline sérique (PAS), les protéines totales (PT) et l'acétylcholine estérase (AchE) et les caractères de production et de reproduction ont été estimées. Des estimations d'héritabilité plus élevées ont été obtenues pour le poids corporel et la longueur des cuisses des lignées sélectionnées. De grands différentiels de sélection ont été obtenus pour le poids corporel et les caractères de conformation corporelle. Les réponses à la sélection ont été les plus importantes dans les lignées sélectionnées. Le nombre d'œufs le plus élevé et le plus significatif a été signalé dans la lignée maternelle sélectionnée. Les lignées sélectionnées ont révélé un poids d'œufs plus élevé. PAS était positivement et significativement corrélé au poids des œufs (r = 0,761, p = 0,001), à la fertilité (r = 0,870, p = 0,0001) et à l'éclosion (r = 0,915, p = 0,0001). Les protéines totales et l'acétylcholine estérase étaient positivement corrélées au nombre d'œufs. Il est donc conclu dans cette étude que la sélection pourrait conduire à une amélioration des caractéristiques de croissance et de reproduction chez le poulet de chair et que PAS, PT et AchE pourraient être utilisés comme marqueurs pour sélectionner le poids et le nombre d'œufs chez le poulet de chair. de population croissante. Cette recherche visait à évaluer l'effet de la réponse de sélection sur les

An online fault detection and diagnosis method of sensors in district heating substations based on long short-term memory network and adaptive threshold selection algorithm

District heating system is an essential energy infrastructure. Due to the harsh working environment, sensors in district heating substations are prone to faults, which will mislead the control strategies of substations. We propose an online fault detection and diagnosis (FDD) method for sensors in substations. The proposed method establishes a Sequence-to-Sequence Prediction Model based on Long Short-Term Memory network with Spatial-Temporal Attention (abbreviated as STALSTM-seq2seq model) for each target sensor. The model can learn normal and evolving data patterns from historical time series of sensors and predict current sensor observation. Then, an adaptive threshold selection algorithm is used to autonomously determine current anomaly threshold based on the prediction error vector of the model. If the prediction error exceeds the anomaly threshold, the sensor malfunctions. Otherwise, the sensor operates normally. Through a comprehensive case study, we demonstrate that the adaptive threshold selection algorithm can fully utilize the forecasting ability of the STALSTM-seq2seq model, and thereby ensure the accuracy of sensor FDD. Overall, the online sensor FDD method achieves an average F1 score of 0.8473. The research shows that the proposed method provides an effective and robust solution to timely and accurately identify sensor faults for district heating substations.

Diversity assessment of a lesser known buffalo population from Central India and its comparative evaluation reveals presence of sufficient genetic variation and absence of selection

Planned breeding and conservation strategies for a lesser-known population require an assessment of complete genetic diversity and population structure analysis in addition to its morphometric characteristics. In the present study, a comparative analysis of the genetic structure of a rare buffalo population, namely Chhattisgarhi, was extensively studied using a panel of FAO-recommended microsatellite markers along with well-established breeds namely Murrah, Nili-Ravi, Gojri, Kalahandi, and Nagpuri. Mode shift analysis indicated the absence of genetic bottleneck in the recent past. Assessment of genetic diversity indices across all loci indicated the presence of sufficient genetic variation within and between populations. Analysis of molecular variance between the six different buffalo populations attributed 19.05% of the variations to between-population differentiation. Cluster analyses using DAPC and Bayesian approach along with the phylogenetic tree based on UPGMA grouped six populations into three groups. The Chhattisgarhi population was revealed to be genetically closer to Nagpuri and Kalahandi populations. The study reveals the presence of sufficient genetic diversity within the Chhattisgarhi population and indicates the absence of a systematic selection program. We suggest improvement and conservation programs should be planned for this breed in the near future through short-term selection.

Transcriptional signature of host shift in the seed beetle Zabrotes subfasciatus.

In phytophagous insects, adaptation to a new host is a dynamic process, in which early and later steps may be underpinned by different features of the insect genome. Here, we tested the hypothesis that early steps of this process are underpinned by a shift in gene expression patterns. We set up a short-term artificial selection experiment (10 generations) for the use of an alternative host (Cicer arietinum) on populations of the bean beetle Zabrotes subfasciatus. Using Illumina sequencing on young adult females, we show the selected populations differ in the expression of genes associated to stimuli, signalling, and developmental processes. Particularly, the "C. arietinum" population shows upregulation of histone methylation genes, which may constitute a strategy for fine-tuning the insect global gene expression network. Using qPCR on body regions, we demonstrated that the "Phaseolus vulgaris" population upregulates the genes polygalacturonase and egalitarian and that the expression of an odorant receptor transcript variant changes over generations. Moreover, in this population we detected the existence of vitellogenin (Vg) variants in both males and females, possibly harbouring canonical reproductive function in females and extracellular unknown functions in males. This study provides the basis for future genomic investigations seeking to shed light on the nature of the proximate mechanisms involved in promoting differential gene expression associated to insect development and adaptation to new hosts.

Leafy localities, longer lives: A cross-sectional and spatial analysis

Are trees good for your health? Using detailed satellite imagery, we estimate the extent of tree coverage at a fine neighborhood level across urban Australia. We then look at the neighborhood-level association between tree canopy cover and mortality. Holding constant socioeconomic status, we find evidence of a strong beneficial relationship. Neighborhoods with more trees have lower levels of mortality, with a 10 percentage point increase in tree cover (about one standard deviation) associated with a reduction in mortality of 11 deaths per 100,000 people (about one eighth of a standard deviation). This association holds for most major causes of death, and is larger for men than for women. Health morbidity is better in areas with more trees, although this relationship is not statistically significant. Analysis of sub-samples does not support the critique that our results are merely driven by short-term selection effects in which healthier people move to tree-lined suburbs. Using standard estimates of the value of a statistical life, the mortality benefit of additional trees substantially exceeds the cost of planting and maintenance. Our findings support the protection and restoration of tree canopy in urban neighborhoods as a means of promoting public health and reducing health inequalities.

A Mouse-Specific Model to Detect Genes under Selection in Tumors.

The mouse is a widely used model organism in cancer research. However, no computational methods exist to identify cancer driver genes in mice due to a lack of labeled training data. To address this knowledge gap, we adapted the GUST (Genes Under Selection in Tumors) model, originally trained on human exomes, to mouse exomes via transfer learning. The resulting tool, called GUST-mouse, can estimate long-term and short-term evolutionary selection in mouse tumors, and distinguish between oncogenes, tumor suppressor genes, and passenger genes using high-throughput sequencing data. We applied GUST-mouse to analyze 65 exomes of mouse primary breast cancer models and 17 exomes of mouse leukemia models. Comparing the predictions between cancer types and between human and mouse tumors revealed common and unique driver genes. The GUST-mouse method is available as an open-source R package on github.

A common garden experiment supports a genetic component underlying the increased resilience of common cockle (Cerastoderma edule) to the parasite Marteilia cochillia.

The common cockle is a valuable bivalve species inhabiting the Atlantic European coasts. The parasite Marteilia cochillia has devastated cockle beds in the southern Galician (NW Spain) rias since 2012. Previous data suggested that cockles from Ría de Arousa acquired some resilience to this parasite through natural selection after consecutive annual marteiliosis outbreaks and candidate markers associated with marteiliosis resilience were identified using population genomics and transcriptomics approaches. Here, a common garden experiment was performed using a naïve stock (from Ría de Muros-Noia) and an affected stock (from Ría de Arousa) to test this hypothesis. Breeders from both stocks were used to produce seed cohorts at hatchery, which were pre-grown in a raft (outdoor nursery stage) and deployed in two shellfish beds affected by marteiliosis in Ría de Arousa (growing-out stage). In both beds, the naïve stock showed high marteiliosis prevalence and was fully depleted in a short period, while the affected stock barely showed evidence of marteiliosis. A set of 45 SNPs putatively associated with marteiliosis resilience were fitted for MassARRAY genotyping to check their role in the differential resilience detected between both stocks. Though no significant differentiation was found between the naïve and the affected stocks with neutral markers, 28 SNPs showed significant divergence between them, suggesting that these SNPs were involved in directional selection during eight generations (to the most) of marteiliosis pressure (long-term selection). Furthermore, signals of selection were also detected in the naïve stock along the marteiliosis outbreak in the growing-out stage (short-term selection) and six SNPs, all shared with the long-term evaluation, showed consistent signals of differentiation according to the infection severity. Some of these SNPs were located within immune genes pertaining to families such as proteasome, ubiquitin, tumor necrosis factor, and glutathione S-transferase. These resilience-associated markers will be useful to recover cockle production in Galicia.

Dropout prediction and decision feedback supported by multi temporal sequences of learning behavior in MOOCs

The temporal sequence of learning behavior is multidimensional and continuous in MOOCs. On the one hand, it supports personalized learning methods, achieves flexible time and space. On the other hand, it also makes MOOCs produce a large number of dropouts and incomplete learning behaviors. Dropout prediction and decision feedback have become an important issue of MOOCs. This study carries out sufficient method design and decision analysis on the dropout trend. Based on a large number of learning behavior instances, we construct a multi behavior type association framework, design dropout prediction model to analyze the temporal sequence of learning behavior, then discuss the corresponding intervention measures, in order to provide adaptive monitoring mechanism for long-term tracking and short-term learning method selection, and enable adaptive decision feedback. the full experiment shows that the designed model might improve the performance of the dropout prediction, which achieves the reliability and feasibility. The whole research can provide key technical solution and decision, which has important theoretical and practical value for dropout research of MOOCs.

Resequencing of a Pekin duck breeding population provides insights into the genomic response to short-term artificial selection.

Short-term, intense artificial selection drives fast phenotypic changes in domestic animals and leaves imprints on their genomes. However, the genetic basis of this selection response is poorly understood. To better address this, we employed the Pekin duck Z2 pure line, in which the breast muscle weight was increased nearly 3-fold after 10 generations of breeding. We denovo assembled a high-quality reference genome of a female Pekin duck of this line (GCA_003850225.1) and identified 8.60 million genetic variants in 119 individuals among 10 generations of the breeding population. We identified 53 selected regions between the first and tenth generations, and 93.8% of the identified variations were enriched in regulatory and noncoding regions. Integrating the selection signatures and genome-wide association approach, we found that 2 regions covering 0.36 Mb containing UTP25 and FBRSL1 were most likely to contribute to breast muscle weight improvement. The major allele frequencies of these 2 loci increased gradually with each generation following the same trend. Additionally, we found that a copy number variation region containing the entire EXOC4 gene could explain 1.9% of the variance in breast muscle weight, indicating that the nervous system may play a role in economic trait improvement. Our study not only provides insights into genomic dynamics under intense artificial selection but also provides resources for genomics-enabled improvements in duck breeding.

Experimental laboratory tests of short-term habitat selection by hatchery-reared juvenile red drum (Sciaenops ocellatus)

Structured estuarine habitats enhance the survival and growth of wild juvenile fishes by mediating density-dependent processes. However, the habitat selection behavior of hatchery-reared fishes upon release to the wild is not well understood. The red drum (Sciaenops ocellatus) is a highly sought-after, estuarine-dependent sportfish and the subject of stock enhancement efforts along the southeastern US Atlantic and Gulf coasts. We investigated the habitat associations of hatchery-reared red drum using experimental mesocosms intended to mimic the habitat types available at or nearby typical release sites in South Carolina, USA. Specifically, we presented juvenile red drum (overall mean ∼ 60 mm standard length) with a choice between salt marsh vegetation (Spartina alterniflora) and oyster shell in the presence and absence of a common estuarine predator (blue crab, Callinectes sapidus). Red drum more commonly associated with oyster shell over Spartina in both the presence and absence of a blue crab predator. Oyster shell habitat tended to be favored even when the location of the free-ranging blue crab predator was taken into account. Our results provide evidence that oyster shells, representative of a healthy restored or natural oyster reef, may be preferentially utilized by hatchery-reared red drum during the period immediately following release. Further investigation of post-release behaviors, including field-based experimental studies focusing on hatchery-reared red drum, is necessary to determine if the patterns of habitat selection behavior documented here have implications for post-release survival or mortality.

The dynamic changes of genes revealed that persistently overexpressed genes drive the evolution of cyflumetofen resistance in Tetranychus cinnabarinus.

Changes in gene expression are associated with the evolution of pesticide resistance in arthropods. In this study, transcriptome sequencing was performed in 3 different resistance levels (low, L; medium, M; and high, H) of cyflumetofen-resistant strain (YN-CyR). A total of 1685 genes, including 97 detoxification enzyme genes, were upregulated in all 3 stages, of which 192 genes, including 11 detoxification enzyme genes, showed a continuous increase in expression level with resistance development (L to H). RNA interference experiments showed that overexpression of 7 genes (CYP392A1, TcGSTd05, CCE06, CYP389A1, TcGSTz01, CCE59, and CYP389C2) is involved in the development of cyflumetofen resistance in Tetranychus cinnabarinus. The recombinant CYP392A1 can effectively metabolize cyflumetofen, while CCE06 can bind and sequester cyflumetofen in vitro. We compared 2 methods for rapid screening of resistance molecular markers, including short-term induction and 1-time high-dose selection. Two detoxification enzyme genes were upregulated in the field susceptible strain (YN-S) by induction with 20% lethal concentration (LC20 ) of cyflumetofen. However, 16 detoxification enzyme genes were upregulated by 1-time selection with LC80 of cyflumetofen. Interestingly, the 16 genes were overexpressed in all 3 resistance stages. These results indicated that 1685 genes that were upregulated at the L stage constituted the basis of cyflumetofen resistance, of which 192 genes in which upregulation continued to increase were the main driving force for the development of resistance. Moreover, the 1-time high-dose selection is an efficient way to rapidly obtain the resistance-related genes that can aid in the development of resistance markers and resistance management in mites.

Contributions of adaptation and purifying selection to SARS-CoV-2 evolution.

Continued evolution and adaptation of SARS-CoV-2 has led to more transmissible and immune-evasive variants with profound impacts on the course of the pandemic. Here I analyze the evolution of the virus over 2.5 years since its emergence and estimate the rates of evolution for synonymous and non-synonymous changes separately for evolution within clades-well-defined monophyletic groups with gradual evolution-and for the pandemic overall. The rate of synonymous mutation is found to be around 6 changes per year. Synonymous rates within variants vary little from variant to variant and are compatible with the overall rate of 7 changes per year (or [Formula: see text] per year and codon). In contrast, the rate at which variants accumulate amino acid changes (non-synonymous mutations) was initially around 12-16 changes per year, but in 2021 and 2022 it dropped to 6-9 changes per year. The overall rate of non-synonymous evolution, that is across variants, is estimated to be about 26 amino acid changes per year (or [Formula: see text] per year and codon). This strong acceleration of the overall rate compared to within clade evolution indicates that the evolutionary process that gave rise to the different variants is qualitatively different from that in typical transmission chains and likely dominated by adaptive evolution. I further quantify the spectrum of mutations and purifying selection in different SARS-CoV-2 proteins and show that the massive global sampling of SARS-CoV-2 is sufficient to estimate site-specific fitness costs across the entire genome. Many accessory proteins evolve under limited evolutionary constraints with little short-term purifying selection. About half of the mutations in other proteins are strongly deleterious.

Species dynamics in natural bacterial communities over multiple rounds of propagation.

Our experimental work illustrates how microbial ecosystems can be shaped by selective pressures over long-term ecological time scales. Natural microbial ecosystems generally consist of various co-existing species, where community composition describes the frequency at which species or types are present. Overall functionality of the system is achieved by interacting species. Upon short-term selection, for instance by transfer to a novel environment, community composition and functionality may change in a process referred to as species sorting. Various factors, such as initial community composition and selective pressures from the environment, may influence this change. Mabisi is a traditional fermented food from Zambia that naturally contains a bacterial community of around twenty unique bacterial types. We used six comparable but different natural bacterial Mabisi communities, each split into five identical replicates, for 16 propagation cycles in a novel, common laboratory environment. Composition of the bacterial communities changed upon propagation. The influence of four main factors on community composition, i.e. initial composition (history), impact of the environment (adaptation), changes due to interaction between species and random processes (chance) in species dynamics, was tested using maximum likelihood ratios. Initial community composition seemed to determine the change in community composition, followed by random processes. Interestingly, we observed convergence at the level of ecosystem functionality, which was measured as profiles of metabolic output. This suggests different combinations of species or types can achieve similar eco-system functionality.

Short-Term Traffic Flow Prediction with Structure-Optimized Deep Belief Network

This paper proposes a structure-optimized deep belief network method for short-term traffic flow forecast, which is used to solve the problems of too simple training data in deep learning short-term traffic flow forecast and random selection of model structure construction parameters. We constructed a deep belief network short-term traffic flow forecast model that can simultaneously train three types of traffic data related to the predicted node traffic volume, enhance the spatiotemporal correlation of predictions, and overcome the shortcomings of too single training data. At the same time, we optimize the short-term traffic flow prediction model structure of the deep belief network; and use the T-PSO algorithm to optimize the hidden layer structure parameters of the prediction model. It can avoid the decrease in the practicability of the model caused by the random selection of the model structure construction parameters. The experimental results show that the method of structure-optimized deep belief network is feasible and effective, and the prediction accuracy is better than the classic deep learning prediction model.

Metabolic Adaptation of a Globally Important Diatom following 700 Generations of Selection under a Warmer Temperature.

Diatoms, accounting for 40% of the marine primary production and 20% of global carbon dioxide fixation, are threatened by the ongoing ocean warming (OW). However, whether and how these ecologically important phytoplankton adapt to OW remains poorly unknown. Here, we experimentally examined the metabolic adaptation of a globally important diatom species Skeletonema dohrnii (S. dohrnii) to OW at two elevated temperatures (24 and 28 °C compared with 20 °C) under short-term (∼300 generations) and long-term (∼700 generations) selection. Both warming levels significantly increased the cell growth rate but decreased the chlorophyll a content. The contents of particulate organic carbon (POC) and particulate organic nitrogen (PON) decreased significantly initially (i.e., until 300 generations) at two temperature treatments but completely recovered after 700 generations of selection, suggesting that S. dohrnii ultimately developed thermal adaptation. Proteomic analysis demonstrated that elevated temperatures upregulated energy metabolism via glycolysis, tricarboxylic acid cycle, and fatty acid oxidation as well as nitrogen acquisition and utilization, which in turn reduced substance storage because of trade-off in the 300th generation, thus decreasing POC and PON. Interestingly, populations at both elevated temperatures exhibited significant proteome plasticity in the 700th generation, as primarily demonstrated by the increased lipid catabolism and glucose accumulation, accounting for the recovery of POC and PON. Changes occurring in cells at the 300th and 700th generations demonstrate that S. dohrnii can adapt to the projected OW, and readjusting the energy metabolism is an important adaptive strategy.