Extra Set Research Articles

Optimal control of a doubly fed induction generator of a wind turbine in cooperation with weak and rigid grids

In this article a control method for a rotor side converter (RSC) of a doubly fed induction generator of a wind turbine is developed. The doubly-fed induction generator (DFIG) system of a wind energy power plant that improves grid symmetrization was applied. The issue of optimal control was treated as an extended linear quadratic regulator (ELQR) having an extra set of exogenous inputs, which are source voltages of the electric grid. No additional knowledge of equations modelling the exogenous inputs was assumed. The proposed method is much more efficient than the currently available linear quadratic control methods. The control objective for a weak grid was to maintain the given value of the voltage module on load terminals and the given value of active power transferred from the stator’s winding. First harmonic components of relevant waveforms were used for this purpose. This task also required the DFIG system to provide reactive power to the grid. In the case of a rigid grid, this reactive power would be too high. Therefore, in this case, it was assumed that the system would supply only part of the required active and reactive power, based on its capabilities. It was required that the voltage and current ratings of the system, mainly the DFIG, were not exceeded. Therefore, the parameters of the network in these difficult failure cases were corrected only partially. The behaviour of the grid in the conditions of failure, as well as the return to the steady state after failure disappearance, were studied.

Уравнения Лакса на супералгебрах Ли

It is demonstrated that the standard construction of Lax equations on Lie algebras can be extended to Lie superalgebras, with the even subspace carrying the usual Lax equations. The extended equations inherit the existence of the canonical trace polynomial integrals of motion. An extra set of integrals exists in the odd subspace, with a nontrivial homological structure of the orbit space. This establishes a curious algebraic link between integrable evolution equations, supersymmetry and the deformation theory.

Automatic hepatic tumor segmentation in intra-operative ultrasound: a supervised deep-learning approach.

Training and evaluation of the performance of a supervised deep-learning model for the segmentation of hepatic tumors from intraoperative US (iUS) images, with the purpose of improving the accuracy of tumor margin assessment during liver surgeries and the detection of lesions during colorectal surgeries. In this retrospective study, a U-Net network was trained with the nnU-Net framework in different configurations for the segmentation of CRLM from iUS. The model was trained on B-mode intraoperative hepatic US images, hand-labeled by an expert clinician. The model was tested on an independent set of similar images. The average age of the study population was 61.9 ± 9.9 years. Ground truth for the test set was provided by a radiologist, and three extra delineation sets were used for the computation of inter-observer variability. The presented model achieved a DSC of 0.84 (), which is comparable to the expert human raters scores. The model segmented hypoechoic and mixed lesions more accurately (DSC of 0.89 and 0.88, respectively) than hyper- and isoechoic ones (DSC of 0.70 and 0.60, respectively) only missing isoechoic or >20mm in diameter (8% of the tumors) lesions. The inclusion of extra margins of probable tumor tissue around the lesions in the training ground truth resulted in lower DSCs of 0.75 (). The model can accurately segment hepatic tumors from iUS images and has the potential to speed up the resection margin definition during surgeries and the detection of lesion in screenings by automating iUS assessment.

Novel method of multiaxis weaving and impact analysis of the ensuing composites

The study aims to develop a novel method of weaving multiaxis preform on conventional weaving machines and analyse impact tolerance of the ensuing composites along with development of finite element analysis model. An extra set of 200 tex carbon fibres, referred as bias fibres, were integrated into carbon fibre plain woven fabric during the weaving stage on a conventional closed reed weaving machine. This was achieved by partial carbon fibre weft insertion at regular intervals from both edges of the fabric. The multiaxis preform was used as ply on top of plain woven carbon fabric plies, each of 160 g per square meter aerial density. The carbon fibre composites were manufactured by infusing bisphenol F epoxy resin as a matrix and 4 plies of carbon fabric by hand layup method. The composites were subjected to low-velocity impact loading at 0.4 m height of 5.12 kg impactor, and force-time history was recorded by the piezoelectric transducer. A comparison of impact loading results showed that the multiaxis woven composites increase impact strength by 18.3% due to the reinforcement of bias fibre. Based on the test results, a finite element analysis model was developed in Ansys v.19 to simulate the stress distribution during impact loading of the multiaxis woven composites.

Consistency-Aware Anchor Pyramid Network for Crowd Localization.

Crowd localization aims to predict the positions of humans in images of crowded scenes. While existing methods have made significant progress, two primary challenges remain: (i) a fixed number of evenly distributed anchors can cause excessive or insufficient predictions across regions in an image with varying crowd densities, and (ii) ranking inconsistency of predictions between the testing and training phases leads to the model being sub-optimal in inference. To address these issues, we propose a Consistency-Aware Anchor Pyramid Network (CAAPN) comprising two key components: an Adaptive Anchor Generator (AAG) and a Localizer with Augmented Matching (LAM). The AAG module adaptively generates anchors based on estimated crowd density in local regions to alleviate the anchor deficiency or excess problem. It also considers the spatial distribution prior to heads for better performance. The LAM module is designed to augment the predictions which are used to optimize the neural network during training by introducing an extra set of target candidates and correctly matching them to the ground truth. The proposed method achieves favorable performance against state-of-the-art approaches on five challenging datasets: ShanghaiTech A and B, UCF-QNRF, JHU-CROWD++, and NWPU-Crowd. The source code and trained models will be released at https://github.com/ucasyan/CAAPN.

Applicability of Fine-Tuned ChatGPT to Automated Essay Scoring

ChatGPT, released in 2022, has garnered attention due to its adaptability through prompt engineering, enabling users to guide its responses. It is important to note that the extent to which users can modify ChatGPT remains limited, as its core embeddings stay unaltered through prompt engineering. Thus, this study aims to evaluate the effectiveness of ChatGPT and its fine-tuned model in essay evaluation compared to human raters. A total of 904 essays from the YELC 2011, on the subject of physical punishment, were selected for this study. Among these, 723 essays were used for fine-tuning ChatGPT, and the remaining 181 were reserved for testing the language model. Additionally, an extra set of 200 essays on different topics, such as driving and medical issues, was included to evaluate the language model’s performance across various themes. Inter-rater reliability indices, including measures like correlation, agreement, Cohen’s kappa, and Krippendorff’s alpha, along with many-facet Rasch measurement analysis, collectively indicated that the current version of ChatGPT (gpt-3.5-turbo-0613) is not yet poised to fully supplant human raters in essay scoring. Nevertheless, through the fine-tuning process, the model demonstrated a significant level of agreement with human raters and exhibited a marked degree of consistency.

Construction of a risk stratification model integrating ctDNA to predict response and survival in neoadjuvant-treated breast cancer

BackgroundThe pathological complete response (pCR) to neoadjuvant chemotherapy (NAC) of breast cancer is closely related to a better prognosis. However, there are no reliable indicators to accurately identify which patients will achieve pCR before surgery, and a model for predicting pCR to NAC is required.MethodsA total of 269 breast cancer patients in Shandong Cancer Hospital and Liaocheng People’s Hospital receiving anthracycline and taxane-based NAC were prospectively enrolled. Expression profiling using a 457 cancer-related gene sequencing panel (DNA sequencing) covering genes recurrently mutated in breast cancer was carried out on 243 formalin-fixed paraffin-embedded tumor biopsies samples before NAC from 243 patients. The unique personalized panel of nine individual somatic mutation genes from the constructed model was used to detect and analyze ctDNA on 216 blood samples. Blood samples were collected at indicated time points including before chemotherapy initiation, after the 1st NAC and before the 2nd NAC cycle, during intermediate evaluation, and prior to surgery. In this study, we characterized the value of gene profile mutation and circulating tumor DNA (ctDNA) in combination with clinical characteristics in the prediction of pCR before surgery and investigated the prognostic prediction. The median follow-up time for survival analysis was 898 days.ResultsFirstly, we constructed a predictive NAC response model including five single nucleotide variant (SNV) mutations (TP53, SETBP1, PIK3CA, NOTCH4 and MSH2) and four copy number variation (CNV) mutations (FOXP1-gain, EGFR-gain, IL7R-gain, and NFKB1A-gain) in the breast tumor, combined with three clinical factors (luminal A, Her2 and Ki67 status). The tumor prediction model showed good discrimination of chemotherapy sensitivity for pCR and non-pCR with an AUC of 0.871 (95% CI, 0.797–0.927) in the training set, 0.771 (95% CI, 0.649–0.883) in the test set, and 0.726 (95% CI, 0.556–0.865) in an extra test set. This tumor prediction model can also effectively predict the prognosis of disease-free survival (DFS) with an AUC of 0.749 at 1 year and 0.830 at 3 years. We further screened the genes from the tumor prediction model to establish a unique personalized panel consisting of 9 individual somatic mutation genes to detect and analyze ctDNA. It was found that ctDNA positivity decreased with the passage of time during NAC, and ctDNA status can predict NAC response and metastasis recurrence. Finally, we constructed the chemotherapy prediction model combined with the tumor prediction model and pretreatment ctDNA levels, which has a better prediction effect of pCR with the AUC value of 0.961.ConclusionsIn this study, we established a chemotherapy predictive model with a non-invasive tool that is built based on genomic features, ctDNA status, as well as clinical characteristics for predicting pCR to recognize the responders and non-responders to NAC, and also predicting prognosis for DFS in breast cancer. Adding pretreatment ctDNA levels to a model containing gene profile mutation and clinical characteristics significantly improves stratification over the clinical variables alone.

The 5th Dimension and its Implications for the String Theory, Conservation of Energy and Heisenberg Uncertainty Principle

In 1919, the mathematician Theodor Kaluza began to play around with Albert Einstein’s formulas for gravity. Out of curiosity, he reworked the equations to see how they‘d look in five, rather than four, dimensions. The exercise yielded an extra set of equations, which were the same as James Clerk Maxwell’s equations for the electromagnetic field. His idea of a fifth dimension had produced a mathematical unification of gravity and electromagnetism. Since Kaluza’s calculations yielded an extra set of equations with Einstein’s formulas for gravity, it’s logical to assume that extending the number of known dimensions to five might also extend E = mc2. This article also visits conservation of mass and energy / mass-energy equivalence, Electromagnetism, Dark Matter, Dark Energy, the Matrix, Antigravitons, Incompleteness Theorems, New Irrationals, Calculus, Celluloid Motion, and Vector-Tensor-Scalar Geometry. And as the text accompanying Figure 5 says, “The counterclockwise rotation does not have to encompass 360 degrees. It can be divided into single degrees – and even arcseconds which are 3,600 times smaller than a degree. Each arcsecond (or tiny part of it) could correspond to a separate dimension and the total dimensions might make up a temporal multiverse that’s reminiscent of string theory, the cosmological framework that says particles are composed of one-dimensional objects called strings (a group of binary digits is called a single-dimensional bit string).” Finally, Heisenberg’s Principle of Indeterminacy - in which the position and the velocity of a particle cannot both be measured exactly, at the same time, even in theory - is revised. Quantum uncertainty is changed into quantum certainty by software containing a new type of irrational number – actually, irrational equations - that’s based on mathematics’ Matrix and topology’s Mobius strip, as well as being interdimensional. Ludwig Wittgenstein, an Austrian-British philosopher, once stated, "The limits of my language mean the limits of my world." Wittgenstein proposed that our understanding of the world is shaped by the language we use to describe it. This concept extends to mathematics, which can be seen as a specialized language. So if a person has no doubt that 1 + 1 = 2, he or she can never arrive at a truly Unified Theory in which only one thing exists. The maths accepted in 2023 is indispensable in enabling advancements in technology and science that have transformed our lives but is incapable of fully describing Unification. Scientific history has shown that it can go partway. Scientists consequently assume it will eventually go all the way.

Multi-user immersive environment for excavator teleoperation in construction

Excavation on jobsites is a collaborative effort, with a spotter serving as an extra set of eyes for the operator to ensure safety. However, current training methods using simulators are limited in immersion and primarily designed for individual operators to practice basic skills. This paper presents a multi-user teleoperation system featured with hybrid-immersive interface and between-user communication. The effectiveness of the system is evaluated through human subject experiments. The study's findings indicate that operators experienced greater immersion, a stronger sense of presence, and improved user interaction when using an immersive interface. This paper has important implications for the future of excavation training, with potential benefits including reduced utility strike damage and the opportunity to investigate human-robot collaboration in jobsites with multiple workers performing various roles within a highly immersive operational environment.

Machine learning-based prediction model of acute kidney injury in patients with acute respiratory distress syndrome

BackgroundAcute kidney injury (AKI) can make cases of acute respiratory distress syndrome (ARDS) more complex, and the combination of the two can significantly worsen the prognosis. Our objective is to utilize machine learning (ML) techniques to construct models that can promptly identify the risk of AKI in ARDS patients.MethodWe obtained data regarding ARDS patients from the Medical Information Mart for Intensive Care III (MIMIC-III) and MIMIC-IV databases. Within the MIMIC-III dataset, we developed 11 ML prediction models. By evaluating various metrics, we visualized the importance of its features using Shapley additive explanations (SHAP). We then created a more concise model using fewer variables, and optimized it using hyperparameter optimization (HPO). The model was validated using the MIMIC-IV dataset.ResultA total of 928 ARDS patients without AKI were included in the analysis from the MIMIC-III dataset, and among them, 179 (19.3%) developed AKI after admission to the intensive care unit (ICU). In the MIMIC-IV dataset, there were 653 ARDS patients included in the analysis, and among them, 237 (36.3%) developed AKI. A total of 43 features were used to build the model. Among all models, eXtreme gradient boosting (XGBoost) performed the best. We used the top 10 features to build a compact model with an area under the curve (AUC) of 0.850, which improved to an AUC of 0.865 after the HPO. In extra validation set, XGBoost_HPO achieved an AUC of 0.854. The accuracy, sensitivity, specificity, positive prediction value (PPV), negative prediction value (NPV), and F1 score of the XGBoost_HPO model on the test set are 0.865, 0.813, 0.877, 0.578, 0.957 and 0.675, respectively. On extra validation set, they are 0.724, 0.789, 0.688, 0.590, 0.851, and 0.675, respectively.ConclusionML algorithms, especially XGBoost, are reliable for predicting AKI in ARDS patients. The compact model maintains excellent predictive ability, and the web-based calculator improves clinical convenience. This provides valuable guidance in identifying AKI in ARDS, leading to improved patient outcomes.

Enhancement of loop mediated isothermal amplification's sensitivity and speed by multiple inner primers for more efficient identification of Vibrio parahaemolyticus

The modified loop-mediated isothermal amplification (LAMP), called multiple hybrid, inner primers (MHP)-LAMP, was developed to enhance the efficiency of the existing LAMP-based assay for Vibrio parahaemolyticus detection. The method was built on a conventional LAMP assay by employing 2 newly designed extra sets of primers to increase the initial binding sites of core primers on the V. parahaemolyticus’s rpoD gene from 8 to 12. With this strategy, the assay detection sensitivity was increased by 10 folds, with the detection limit (DL) approaching 100 copies of purified target genomic DNA (gDNA) as analyzed by real-time turbidity measurement and gel electrophoresis. The MHP also accelerated the rate of DNA amplification by 30%, rendering the assay faster. The MHP-LAMP assay did not cross- react with other pathogens, indicating that it was highly specific for V. parahaemolyticus detection. Whilst V. parahaemolyticus was used as a study model herein, our idea of using MHP to maximize assay sensitivity and speed is considered as a universal strategy that can be applied to enhance efficiency of LAMP-based assays for detecting any DNA and RNA of interest.•The strategy of using multiple hybrid, inner primers (MHP) to enhance LAMP assay's efficiency was demonstrated with success.•The MHP enhanced the sensitivity and speed of the existing LAMP assay, designed to detect V. parahaemolyticus, by 10 times and 30%, respectively.•The proposed strategy can be applied to boost up any other LAMP-based assay's diagnostic performance.

Spatial+: A new cross-validation method to evaluate geospatial machine learning models

Random cross-validation (CV) is often used to evaluate geospatial machine learning models, particularly when a limited amount of sample data are available, and collecting an extra test set is unfeasible. However, the prediction locations can be substantially different from the available sample, leading to over-optimistic evaluation results. This has fostered the development of spatial CV methods. Yet these methods only focus on spatial autocorrelation and cannot sufficiently guarantee that the validation subset is a good proxy of the test set with significant differences. In this paper, we propose the spatial+ cross-validation (SP-CV) method. This method, which considers both the geographic and feature spaces, is composed of two stages. The first stage addresses spatial autocorrelation issues by using agglomerative hierarchical clustering to divide the available sample into blocks. The second stage deals with multiple sources of differences. It uses cluster ensembles to split the blocks into training and validation folds based on the locations of the sample data and the values of the covariates and target variable. The proposed method is compared against random and block CV methods in a series of experiments with Amazon basin above ground biomass and California houseprice datasets. Our results show that SP-CV provided the smallest error differences with respect to the reference error. This means that SP-CV produced more representative splits and led to more reliable model evaluations. It suggests that a reliable model evaluation requires to consider both the geographic and the feature spaces in a comprehensive manner.

O-302 Triploid conceptions are predominantly caused by female meiosis II errors and their risk increases with advancing maternal age

Abstract Study question What are the incidence and origin of ploidy abnormalities in embryos derived from 2 pronuclei (2PN) zygotes, and is their risk linked to maternal age? Summary answer While embryo haploidy is usually due to male meiosis errors, triploidy is mainly caused by female meiosis II errors, whose incidence increases with maternal age. What is known already Normal fertilization is denoted by the appearance of 2PN 16-18 hours after insemination. Deviation from 2PN is considered evidence of abnormal fertilization and ploidy anomalies. Nonetheless, even 2PN embryos can be diagnosed with ploidy abnormalities during preimplantation genetic testing (PGT). The incidence of 3PN has already been linked to advanced maternal age. However, no conclusive and well-powered studies have yet investigated the incidence, origin, and maternal age correlation of genetically-diagnosed ploidy abnormalities. A targeted NGS-based approach that simultaneously analyzes copy number and genotyping data provides a comprehensive ploidy status assessment and helps address numerous basic and clinical research questions. Study design, size, duration Retrospective study including 96,660 trophectoderm biopsies analyzed between 2020 and 2022 using a targetedNGS-based PGT platform. A total of 1,063 embryos carrying haploidy or triploidy were used to investigate maternal age correlation, and parental/meiotic origin of the anomaly. Of these, 57 embryos were from concomitant PGT-M (PGT for monogenic disorders) cycles where also parental DNA was available. These trios (embryo-parents) were used to independently estimate parental/meiotic contribution to the ploidy abnormality and genome-wide recombination events. Participants/materials, setting, methods Targeted-NGS’s accuracy in genotyping and copy number (CN) detection were previously validated on triploid cell lines and multifocal blastocyst biopsies with known ploidy status derived from abnormally fertilized oocytes. Parental and meiotic origins were estimated using two independent approaches. First, they were inferred using sex chromosomes CN from 1,063 embryos with altered ploidy status. Secondly, they were directly calculated using genotyping data from 57 trios (embryo and parents) from PGT-M cycles. Main results and the role of chance The prevalence of ploidy abnormalities in 2PN-derived embryos was 1.1% (n = 1,063/96,660), with triploids accounting for 83.0% (n = 882/1,063) and haploids for 17.0% (n = 181/1,063). Remarkably, the incidence of ploidy anomalies is positively correlated with maternal age (OR = 1.046 per year; p < 0.001). Triploidy showed strong correlation with age (OR = 1.059 per year; p < 0.001), while haploidy did not (OR = 0.96 per year; p = 0.1). Based on sex chromosomes CN analysis, the extra haploid set of triploid embryos was almost completely of maternal origin (94.6%; 95%CI:93.0-96.1), with male errors accounting for only 5.4% (95%CI:4.0-7.1). On the other hand, haploid embryos were the result of paternal errors in 97.8% of cases (95%CI:94.4-99.4), with the missing haploid set being of maternal origin in the remaining 2.2% (95%CI:0.6-5.6). In terms of triploidy’s meiotic origin, two-thirds of the errors occurred during MII (95%CI:63.4-69.8), while one-third occurred during MI (95%CI:30.2-36.5). An independent method using genotyping data of 57 trios confirmed the predominance of paternal error in haploidy (n = 12/14) and the exclusively maternal origin of all embryonic triploidies (n = 43/43). The extra haploid set resulted from an error during MI in 27.9% (n = 12/43) and during MII in 72.1% (n = 31/43) of cases. Interestingly, 16.3% of triploids (n = 7/43) showed no genome-wide recombination events. Limitations, reasons for caution Despite the uniquely large sample size, the inference based on sex chromosomes CN suffers the limitations of the modeling assumptions (independence between parental and meiotic errors), which require further validation. Due to the low prevalence of paternal errors, clinical correlation with male factor could not be investigated with sufficient power. Wider implications of the findings Thanks to the exceedingly high sample size, this is the first study to reveal an increased risk of triploid conception with advancing female age (76% higher at age 40 than at age 30), providing meaningful insights for patients counseling. Importantly, the meiotic origin of ploidy anomalies in embryos were unveiled. Trial registration number not applicable

A multi-center validation study on the discrimination of Legionella pneumophila sg.1, Legionella pneumophila sg. 2-15 and Legionella non-pneumophila isolates from water by FT-IR spectroscopy.

This study developed and validated a method, based on the coupling of Fourier-transform infrared spectroscopy (FT-IR) and machine learning, for the automated serotyping of Legionella pneumophila serogroup 1, Legionella pneumophila serogroups 2-15 as well as their successful discrimination from Legionella non-pneumophila. As Legionella presents significant intra- and inter-species heterogeneities, careful data validation strategies were applied to minimize late-stage performance variations of the method across a large microbial population. A total of 244 isolates were analyzed. In details, the method was validated with a multi-centric approach with isolates from Italian thermal and drinking water (n = 82) as well as with samples from German, Italian, French, and British collections (n = 162). Specifically, robustness of the method was verified over the time-span of 1 year with multiple operators and two different FT-IR instruments located in Italy and Germany. Moreover, different production procedures for the solid culture medium (in-house or commercial) and different culture conditions (with and without 2.5% CO2) were tested. The method achieved an overall accuracy of 100, 98.5, and 93.9% on the Italian test set of Legionella, an independent batch of Legionella from multiple European culture collections, and an extra set of rare Legionella non-pneumophila, respectively.

A lightweight network for portable fry counting devices

Estimating the number of fries plays a critical role in the maintenance of fish breeding, transportation, and the preservation of marine resources in aquaculture. Generally speaking, statistics are recorded manually by fishers and government units. Manual recording is time-consuming and increases the workload of fishers. Compared with traditional physical shunt devices, visual-based algorithms have benefits such as non-restriction of labors, minimal equipment installation, and maintenance costs. However, these methods generally come with massive calculations and model parameters, or poor abilities of aggregation handles and counting precision. This paper proposes a fry counting method named MSENet for portable fry counting devices. Firstly, the lightweight network is designed with simpler parameters (Params: 139.46 kB) for portable embedding. The visualized single-channel fry density maps are predicted by feeding the original images and the number of fries is calculated through integration. Then, the Squeeze-and-Excitation block is utilized to strengthen the features of weighty channels. The model training is refined by hyperparameter studies, the shortened preparation stage enhances the portability. What is more, a fry counting dataset NCAUF and an extra set NCAUF-EX are built for verifications of network generalization. The results demonstrate that the lightweight MSENet outperforms in fry counting with higher precision and competently solves the issue of fry aggregation (MAE: 3.33). The source code and pre-trained models are available at: https://github.com/vranlee/MSENet.

Uni3DA: Universal 3D Domain Adaptation for Object Recognition

Traditional 3D point cloud classification tasks focus on training a classifier in the closed-set scenario, where training and test data have the same label set and the same data distribution. In this work, we focus on a more challenging and realistic scenario in 3D point cloud classification task: universal domain adaptation (UniDA), where 1) data distributions for training and test data are different; and 2) for given label sets of training data and test data, they may have the shared classes and keep the private classes respectively, introducing an extra label set discrepancy. To solve UniDA problem, researchers have designed many methods based on 2D image datasets. However, due to the difficulty in capturing discriminative local geometric structures brought by the unordered and irregular 3D point cloud data, we cannot directly deploy the existing methods based on 2D image datasets to the 3D scenarios. To address UniDA in 3D scenarios, we develop a 3D universal domain adaptation framework, which consists of three modules: Self-Constructed Geometric (SCG) module, Local-to-Global Hypersphere Reasoning (LGHR) module and Self-Supervised Boundary Adaptation (SBA) module. SCG and LGHR generate the discriminative representation, which is used to acquire domain-invariant knowledge for training and test data. SBA is designed to automatically recognize whether a given label is from the shared label set or private label set, and adapts training and test data from the shared label set. To our best knowledge, this work is the first exploration of UniDA for 3D scenarios. Extensive experiments on public 3D point cloud datasets verify that the proposed method outperforms the existing UniDA methods.

Excessive and asymmetrical removal of heterozygous sites by maxSH biases downstream population genetic inference: Implications for hybridization between two primroses

AbstractTechniques of reduced‐representation sequencing (RRS) have revolutionized ecological and evolutionary genomics studies. Precise establishment of orthologs is a critical challenge for RRS, especially when a reference genome is absent. The proportion of shared heterozygous sites across samples is an alternative criterion for filtering paralogs. In the prevailing pipeline for variant calling of RRS data – PYRAD/IPYRAD, maxSH is an often overlooked parameter with implications to detecting and filtering paralogs according to shared heterozygosity. Using empirical genotyping by sequencing data of two primroses (Primula alpicola Stapf and Primula florindae Ward) and their putative hybrids, and extra data sets of Californian golden cup oaks, we explore the impact of maxSH on filtering paralogs and further downstream analyses. Our study sheds light on the simultaneous validity and risk of filtering paralogs using maxSH, and its significant effects on downstream analyses of outlier detection, population assignment, and demographic modeling, emphasizing the importance of attention to detail during bioinformatic processes. The mutual confirmation between results of population assignment and demographic modeling in this study suggested maxSH = 0.10 has a potentially excessive and asymmetrical effect on the removal of truly shared heterozygous sites as paralogs. These results indicate that hybridization origin hypotheses of putative hybrids represented by results with maxSH = 0.25 and 0.50 are more credible. In conclusion, we revealed the critical hazard of paralogs filtration according to sharing heterozygosity at first, so that we propose to use specific protocols, rather than maxSH, to filter potential paralogs for closely related lineages.

Improving fine-tuning of self-supervised models with Contrastive Initialization

Self-supervised learning (SSL) has achieved remarkable performance in pre-training the models that can be further used in downstream tasks via fine-tuning. However, these self-supervised models may not capture meaningful semantic information since the images belonging to the same class are often regarded as negative pairs in the contrastive loss. Consequently, the images of the same class are often located far away from each other in the learned feature space, which would inevitably hamper the fine-tuning process. To address this issue, we seek to explicitly enhance the semantic relation among instances on the targeted downstream task and provide a better initialization for the subsequent fine-tuning. To this end, we propose a Contrastive Initialization (COIN) method that breaks the standard fine-tuning pipeline by introducing an extra class-aware initialization stage before fine-tuning. Specifically, we exploit a supervised contrastive loss to increase inter-class discrepancy and intra-class compactness of features on the target dataset. In this way, self-supervised models can be easily trained to discriminate instances of different classes during the final fine-tuning stage. Extensive experiments show that, with the enriched semantics, our COIN significantly outperforms existing methods without introducing extra training cost and sets new state-of-the-arts on multiple downstream tasks. For example, compared with the baseline method, our COIN improves the accuracy by 5% on ImageNet-20 and 2.57% on CIFAR100, respectively.

Sensitivity Analysis of Krasovskii Passivity-Based Controllers

In the domain of passivity theory, there were major contributions in the last decade, the most recent notion of passivity being the so-called Krasovskii passivity. This framework offers the possibility of designing a controller which ensures the passivity of the resulting closed-loop system. The current paper proposes a solution to design the parameters of Krasovskii passivity-based controllers (K-PBCs) in order to ensure small sensitivity of the closed-loop systems. As such, after the initial construction of the passivity output, the controller parameters are designed in order to impose the dominant eigenvalues of the Jacobian of the resulting closed-loop system with the smallest deviation around the given forced equilibrium point which is, additionally, smaller than a prescribed stability margin. The resulting optimization problem is non-convex by nature, and a metaheuristic approach is proposed to design these parameters. Moreover, in order to impose an extra set of performances, the control system contains an outer loop where dynamical path planning is used to impose the additional requirements. All mentioned results are developed for processes modeled as bilinear systems. In order to illustrate the proposed control method, a numerical example consisting of a single-ended primary inductor DC-DC converter (SEPIC) process is presented.

Insights into the protective capacity of human dental pulp stem cells and its secretome in cisplatin-induced nephrotoxicity: effects on oxidative stress and histological changes.

Acute renal injury (AKI) is a major limiting factor for cisplatin administration. Recent evidence suggests the potential contribution of mesenchymal stem cells (MSCs) to rehabilitation from several disorders via both direct and indirect routes. Thus, the present study aimed, for the first time, to explore and compare the reno-protective potential of human dental pulp-derived stem cells (hDPSCs) vs. hDPSC-conditioned medium (hDPSC-CM) in recovery of impaired kidney tissues in a rat animal model of cisplatin-induced AKI. AKI was induced via cisplatin injection (n=36). One day after, 24 rats were treated with either hDPSCs or hDPSC-CM (n=12). An extra set of rats (n=12) served as sham group. On days 2 or 7 (n=6), rats were humanly sacrificed for further analysis. Renal injury was explored via measuring serum creatinine and BUN. Renal level of oxidative stress was assessed by determining malondialdehyde, and enzymatic activities of superoxide dismutase and catalase. Renal histopathological changes were scored for comparison among different experimental groups. A single dose of cisplatin resulted in considerable renal dysfunction and oxidative stress. Treatment with hDPSCs or hDPSC-CM resulted in significantly restored renal function, reduced level of oxidative stress, and improved histopathological manifestations. Furthermore, as compared to hDPSC-CM, administration of hDPSCs led to superior results in AKI-induced animals. The current study described the first comparative evidence of reno-protective potential of hDPSCs and their CM against cisplatin-induced nephrotoxicity in an AKI rat model, proposing them as useful adjunctive therapy in AKI. Yet, future explorations are still needed.