Forecasting the strength of preplaced aggregate concrete using interpretable machine learning approaches.

Preplaced aggregate concrete (PAC) also known as two-stage concrete (TSC) is widely used in construction engineering for various applications. To produce PAC, a mixture of Portland cement, sand, and admixtures is injected into a mold subsequent to the deposition of coarse aggregate. This process complicates the prediction of compressive strength (CS), demanding thorough investigation. Consequently, the emphasis of this study is on enhancing the comprehension of PAC compressive strength using machine learning models. Thirteen models are evaluated with 261 data points and eleven input variables. The result depicts that xgboost demonstrates exceptional accuracy with a correlation coefficient of 0.9791 and a normalized coefficient of determination (R2) of 0.9583. Moreover, Gradient boosting (GB) and Cat boost (CB) also perform well due to its robust performance. In addition, Adaboost, Voting regressor, and Random forest yield precise predictions with low mean absolute error (MAE) and root mean square error (RMSE) values. The sensitivity analysis (SA) reveals the significant impact of key input parameters on overall model sensitivity. Notably, gravel takes the lead with a substantial 44.7% contribution, followed by sand at 19.5%, cement at 15.6%, and Fly ash and GGBS at 5.9% and 5.1%, respectively. The best fit model i.e., XG-Boost model, was employed for SHAP analysis to assess the relative importance of contributing attributes and optimize input variables. The SHAP analysis unveiled the water-to-binder (W/B) ratio, superplasticizer, and gravel as the most significant factors influencing the CS of PAC. Furthermore, graphical user interface (GUI) have been developed for practical applications in predicting concrete strength. This simplifies the process and offers a valuable tool for leveraging the model's potential in the field of civil engineering. This comprehensive evaluation provides valuable insights to researchers and practitioners, empowering them to make informed choices in predicting PAC compressive strength in construction projects. By enhancing the reliability and applicability of predictive models, this study contributes to the field of preplaced aggregate concrete strength prediction.

[Position paper of the Berlin-Brandenburg Surgical Society-On the future of surgical further training]. / Positionspapier der Berlin-Brandenburgischen Chirurgischen Gesellschaft ­ zur Zukunft der chirurgischen Weiterbildung.

BACKGROUND: Surgical further training faces the challenging task of reconciling technological advancements and patient safety, particularly in the context of the planned hospital reform. Additionally, the generation shift and evolving expectations of Generations Y and Z in the workplace present further challenges. In response to these demands, the Berlin-Brandenburg Surgical Society (Berlin-Brandenburgische Chirurgische Gesellschaft, BCG) initiated a structured discussion and developed a position paper during the Neuhardenberg talks (Neuhardenberger Gespräche). METHODOLOGY: Within the framework of the Neuhardenberg talks, four sessions with keynote presentations and discussions took place. Based on the main discussion points, theses and positions were subsequently formulated and digitally voted on. RESULTS: The results reveal a clear consensus favoring flexible working hours models, earlier specialization options and the integration of external rotations in surgical further training. Regarding talent acquisition and early recruitment of residents, there was a clear consensus supporting the promotion of employee engagement and structured early recruitment of students. There was unanimous agreement on the introduction of training associations as an effective means to ensure high-quality surgical further training. DISCUSSION: One of the central points in the discussions was that high-quality surgical further training will only be achievable within training associations, especially given the impending hospital reform. The BCG plans to develop a modular further training association to make surgical further training in Berlin/Brandenburg fit for the future.

Intersectional race-gender stereotypes in natural language.

How are Asian and Black men and women stereotyped? Research from the gendered race and stereotype content perspectives has produced mixed empirical findings. Using BERT models pre-trained on English language books, news articles, Wikipedia, Reddit and Twitter, with a new method for measuring propositions in natural language (the Fill-Mask Association Test, FMAT), we explored the gender (masculinity-femininity), physical strength, warmth and competence contents of stereotypes about Asian and Black men and women. We find that Asian men (but not women) are stereotyped as less masculine and less moral/trustworthy than Black men. Compared to Black men and Black women, respectively, both Asian men and Asian women are stereotyped as less muscular/athletic and less assertive/dominant, but more sociable/friendly and more capable/intelligent. These findings suggest that Asian and Black stereotypes in natural language have multifaceted contents and gender nuances, requiring a balanced view integrating the gender schema theory and the stereotype content model. Exploring their semantic representations as propositions in large language models, this research reveals how intersectional race-gender stereotypes are naturally expressed in real life.

Using an interpretable deep learning model for the prediction of riverine suspended sediment load.

The prediction of suspended sediment load (SSL) within riverine systems is critical to understanding the watershed's hydrology. Therefore, the novelty of our research is developing an interpretable (explainable) model based on deep learning (DL) and Shapley Additive ExPlanations (SHAP) interpretation technique for prediction of SSL in the riverine systems. This paper investigates the abilities of four DL models, including dense deep neural networks (DDNN), long short-term memory (LSTM), gated recurrent unit (GRU), and simple recurrent neural network (RNN) models for the prediction of daily SSL using river discharge and rainfall data at a daily time scale in the Taleghan River watershed, northwestern Tehran, Iran. The performance of models was evaluated by using several quantitative and graphical criteria. The effect of parameter settings on the performance of deep models on SSL prediction was also investigated. The optimal optimization algorithms, maximum iteration (MI), and batch size (BC) were obtained for modeling daily SSL, and structure of the model impact on prediction remarkably. The comparison of prediction accuracy of the models illustrated that DDNN (with R2 = 0.96, RMSE = 333.46) outperformed LSTM (R2 = 0.75, RMSE = 786.20), GRU (R2 = 0.73, RMSE = 825.67), and simple RNN (R2 = 0.78, RMSE = 741.45). Furthermore, the Taylor diagram confirmed that DDNN has the highest performance among other models. Interpretation techniques can address the black-box nature of models, and here, SHAP was applied to develop an interpretable DL model to interpret of DL model's output. The results of SHAP showed that river discharge has the strongest impact on the model's output in estimating SSL. Overall, we conclude that DL models have great potential in watersheds to predict SSL. Therefore, different interpretation techniques as tools to interpret DL model's output (DL model is as black-box model) are recommended in future research.

Exploiting protein language models for the precise classification of ion channels and ion transporters.

This study introduces TooT-PLM-ionCT, a comprehensive framework that consolidates three distinct systems, each meticulously tailored for one of the following tasks: distinguishing ion channels (ICs) from membrane proteins (MPs), segregating ion transporters (ITs) from MPs, and differentiating ICs from ITs. Drawing upon the strengths of six Protein Language Models (PLMs)-ProtBERT, ProtBERT-BFD, ESM-1b, ESM-2 (650M parameters), and ESM-2 (15B parameters), TooT-PLM-ionCT employs a combination of traditional classifiers and deep learning models for nuanced protein classification. Originally validated on an existing dataset by previous researchers, our systems demonstrated superior performance in identifying ITs from MPs and distinguishing ICs from ITs, with the IC-MP discrimination achieving state-of-the-art results. In light of recommendations for additional validation, we introduced a new dataset, significantly enhancing the robustness and generalization of our models across bioinformatics challenges. This new evaluation underscored the effectiveness of TooT-PLM-ionCT in adapting to novel data while maintaining high classification accuracy. Furthermore, this study explores critical factors affecting classification accuracy, such as dataset balancing, the impact of using frozen versus fine-tuned PLM representations, and the variance between half and full precision in floating-point computations. To facilitate broader application and accessibility, a web server (https://tootsuite.encs.concordia.ca/service/TooT-PLM-ionCT) has been developed, allowing users to evaluate unknown protein sequences through our specialized systems for IC-MP, IT-MP, and IC-IT classification tasks.

Aquatic life criteria of hydrothermal liquefaction wastewater via ecotoxicity test and modeling.

Wastewater resulting from hydrothermal liquefaction (HTL-AP) of biowaste is gaining attention as an emerging hazardous material. However, there is a lack of specific and systematic ecotoxicity studies on HTL-AP. This study addresses this gap by conducting acute toxicity tests on HTL-AP using typical aquatic species and integrating these results with predicted toxicity values from interspecies correlation estimation models to establish aquatic life criteria. HTL-AP exhibited significant toxicity with LC50 of 956.12-3645.4 mg/L, but demonstrated moderate toxicity compared to common freshwater pollutants like commercial microbicides, personal care products, and insect repellents. The resulting hazardous concentration for 5 % of species (HC5), the criterion maximum concentration, and the short-term water quality criteria for aquatic were 506.0, 253.0, and 168.7 mg/L, respectively. Notably, certain organisms like Misgurnus anguillicaudatus and Cipangopaludina chinensis showed high tolerance to HTL-AP, likely due to their metabolic capabilities on HTL-AP components. The significant decrease in HC5 values for some HTL-AP substances compared to pure compounds could indicate the synergistic inhibition effects among HTL-AP compositions. Furthermore, according to the established criteria, HTL-AP required significantly less diluted water (13 t) than carbendazim (1009 t) to achieve biosafety, indicating a safer release. This research establishes a preliminary water quality criterion for HTL-AP, offering a valuable reference for risk assessment and prediction in the utilization of HTL-AP within environmental contexts.

Unlocking the potential of generative AI in drug discovery.

Artificial intelligence (AI) is revolutionizing drug discovery by enhancing precision, reducing timelines and costs, and enabling AI-driven computer-aided drug design. This review focuses on recent advancements in deep generative models (DGMs) for de novo drug design, exploring diverse algorithms and their profound impact. It critically analyses the challenges that are intricately interwoven into these technologies, proposing strategies to unlock their full potential. It features case studies of both successes and failures in advancing drugs to clinical trials with AI assistance. Last, it outlines a forward-looking plan for optimizing DGMs in de novo drug design, thereby fostering faster and more cost-effective drug development.

Effects of Nanomolar Methylmercury on developing Human Neural Stem Cells and Zebrafish Embryo.

Exposure to mercury and its organic form methylmercury (MeHg), is of great concern for the developing nervous system. Despite available literature on MeHg neurotoxicity, there is still uncertainty about its mechanisms of action and the doses that trigger developmental effects. Our study combines two alternative methodologies, the human neural stem cells (NSC) and the zebrafish (ZF) embryo, to address the neurotoxic effects of early exposure to nanomolar concentrations of MeHg. Our results show linear or nonmonotonic (hormetic) responses depending on studied parameters. In ZF, we observed a hormetic response in locomotion and larval rotation, but a concentration-dependent response for sensory organ size and habituation. We also observed a possible delayed response as MeHg had greater effects on larval activity at 5 days than at 24 hours. In NSC cells, some parameters show a clear dose dependence, such as increased apoptosis and differentiation to glial cells or decreased neuronal precursors; while others show a hormetic response: neuronal differentiation or cell proliferation. This study shows that the ZF model was more susceptible than NSC to MeHg neurotoxicity. The combination of different models has improved the understanding of the underlying mechanisms of toxicity and possible compensatory mechanisms at the cellular and organismal level.

More than a number: Incorporating the aged phenotype to improve in vitro and in vivo modeling of neurodegenerative disease.

Age is the number one risk factor for developing a neurodegenerative disease (ND), such as Alzheimer's disease (AD) or Parkinson's disease (PD). With our rapidly ageing world population, there will be an increased burden of ND and need for disease-modifying treatments. Currently, however, translation of research from bench to bedside in NDs is poor. This may be due, at least in part, to the failure to account for the potential effect of ageing in preclinical modelling of NDs. While ageing can impact upon physiological response in multiple ways, only a limited number of preclinical studies of ND have incorporated ageing as a factor of interest. Here, we evaluate the aged phenotype and highlight the critical, but unmet, need to incorporate aspects of this phenotype into both the in vitro and in vivo models used in ND research. Given technological advances in the field over the past several years, we discuss how these could be harnessed to create novel models of ND that more readily incorporate aspects of the aged phenotype. This includes a recently described in vitro panel of ageing markers, which could help lead to more standardised models and improve reproducibility across studies. Importantly, we cannot assume that young cells or animals yield the same responses as seen in the context of ageing; thus, an improved understanding of the biology of ageing, and how to appropriately incorporate this into the modelling of ND, will ensure the best chance for successful translation of new therapies to the aged patient.

Drugging the undruggable: Advances in targeting KRAS signaling in solid tumors.

Cancer remains the leading cause of global mortality, prompting a paradigm shift in its treatment and outcomes with the advent of targeted therapies. Among the most prevalent mutations in RAS-driven cancers, Kirsten rat sarcoma viral oncogene homolog (KRAS) mutations account for approximately 86% of cases worldwide, particularly in lung, pancreatic, and colon cancers, contributing to poor prognosis and reduced overall survival. Despite numerous efforts to understand the biology of KRAS mutants and their pivotal role in cancer development, the lack of well-defined drug-binding pockets has deemed KRAS an "undruggable" therapeutic target, presenting significant challenges for researchers and clinicians alike. Through significant biochemical and technological advances, the last decade has witnessed promising breakthroughs in targeted therapies for KRAS-mutated lung, colon, and pancreatic cancers, marking a critical turning point in the field. In this chapter, we provide an overview of the characteristics of KRAS mutations across various solid tumors, highlighting ongoing cutting-edge research on the immune microenvironment, the development of KRAS-driven mice models, and the recent progress in the exploration of specific KRAS mutant-targeted therapeutic approaches. By comprehensive understanding of the intricacies of KRAS signaling in solid tumors and the latest therapeutic developments, this chapter will shed light on the potential for novel therapeutic strategies to combat KRAS-driven tumors and improve patient outcomes.

Improving Patient Information and Enhanced Consent in Urology: The Impact of Simulation and Multimedia Tools. A Systematic Literature Review from the European Association of Urology Patient Office.

BACKGROUND AND OBJECTIVE: Discussions surrounding urological diagnoses and planned procedures can be challenging, and patients might experience difficulty in understanding the medical language, even when shown radiological imaging or drawings. With the introduction of virtual reality and simulation, informed consent could be enhanced by audiovisual content and interactive platforms. Our aim was to assess the role of enhanced consent in the field of urology. METHODS: A systematic review of the literature was performed in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines, using informed consent, simulation, and virtual reality in urology as the search terms. All original articles were screened. KEY FINDINGS AND LIMITATIONS: Thirteen original studies were included in the review. The overall quality of these studies was deemed good according to the Newcastle-Ottawa Scale. The studies analysed the application of different modalities for enhanced consent: 3D printed or digital models, audio visual multimedia contents, virtual simulation of procedures and interactive navigable apps. Published studies agreed upon a significantly improved effect on patient understanding of the diagnosis, including basic anatomical details, and surgery-related issues such as the aim, steps and the risks connected to the planned intervention. Patient satisfaction was unanimously reported as improved as a result of enhanced consent. CONCLUSIONS AND CLINICAL IMPLICATIONS: Simulation and multimedia tools are extremely valuable for improving patients' understanding of and satisfaction with urological procedures. Widespread application of enhanced consent would represent a milestone for patient-urologist communication. PATIENT SUMMARY: Several multimedia tools can be used to improve patients' understanding of urological conditions and procedures, such as simulation and models. Use of these tools for preoperative discussion enhances knowledge and patient satisfaction, resulting in more realistic patient expectations and better informed consent.

Exploring the influence of invasive weed biochar on the sorption and dissipation dynamics of imazethapyr in sandy loam soil.

The management of invasive weeds on both arable and non-arable land is a vast challenge. Converting these invasive weeds into biochar and using them to control the fate of herbicides in soil could be an effective strategy within the concept of turning waste into a wealth product. In this study, the fate of imazethapyr (IMZ), a commonly used herbicide in various crops, was investigated by introducing such weeds as biochar, i.e., Parthenium hysterophorus (PB) and Lantana camara (LB) in sandy loam soil. In terms of kinetics, the pseudo-second order (PSO) model provided the best fit for both biochar-mixed soils. More IMZ was sorbed onto LB-mixed soil compared to PB-mixed soil. When compared to the control (no biochar), both PB and LB biochars (at concentrations of 0.2% and 0.5%) increased IMZ adsorption, although the extent of this effect varied depending on the dosage and type of biochar. The Freundlich adsorption isotherm provided a satisfactory explanation for IMZ adsorption in soil/soil mixed with biochar, with the adsorption process exhibiting high nonlinearity. The values of Gibb's free energy change (ΔG) were negative for both adsorption and desorption in soil/soil mixed with biochar, indicating that sorption was exothermic and spontaneous. Both types of biochar significantly affect IMZ dissipation, with higher degradation observed in LB-amended soil compared to PB-amended soil. Hence, the findings suggest that the preparation of biochar from invasive weeds and its utilization for managing the fate of herbicides can effectively reduce the residual toxicity of IMZ in treated agroecosystems in tropical and subtropical regions.

A natural language model to automate scoring of autobiographical memories.

Biases in the retrieval of personal, autobiographical memories are a core feature of multiple mental health disorders, and are associated with poor clinical prognosis. However, current assessments of memory bias are either reliant on human scoring, restricting their administration in clinical settings, or when computerized, are only able to identify one memory type. Here, we developed a natural language model able to classify text-based memories as one of five different autobiographical memory types (specific, categoric, extended, semantic associate, omission), allowing easy assessment of a wider range of memory biases, including reduced memory specificity and impaired memory flexibility. Our model was trained on 17,632 text-based, human-scored memories obtained from individuals with and without experience of memory bias and mental health challenges, which was then tested on a dataset of 5880 memories. We used 20-fold cross-validation setup, and the model was fine-tuned over BERT. Relative to benchmarking and an existing support vector model, our model achieved high accuracy (95.7%) and precision (91.0%). We provide an open-source version of the model which is able to be used without further coding, by those with no coding experience, to facilitate the assessment of autobiographical memory bias in clinical settings, and aid implementation of memory-based interventions within treatment services.

Improvements in viral gene annotation using large language models and soft alignments.

BACKGROUND: The annotation of protein sequences in public databases has long posed a challenge in molecular biology. This issue is particularly acute for viral proteins, which demonstrate limited homology to known proteins when using alignment, k-mer, or profile-based homology search approaches. A novel methodology employing Large Language Models (LLMs) addresses this methodological challenge by annotating protein sequences based on embeddings. RESULTS: Central to our contribution is the soft alignment algorithm, drawing from traditional protein alignment but leveraging embedding similarity at the amino acid level to bypass the need for conventional scoring matrices. This method not only surpasses pooled embedding-based models in efficiency but also in interpretability, enabling users to easily trace homologous amino acids and delve deeper into the alignments. Far from being a black box, our approach provides transparent, BLAST-like alignment visualizations, combining traditional biological research with AI advancements to elevate protein annotation through embedding-based analysis while ensuring interpretability. Tests using the Virus Orthologous Groups and ViralZone protein databases indicated that the novel soft alignment approach recognized and annotated sequences that both blastp and pooling-based methods, which are commonly used for sequence annotation, failed to detect. CONCLUSION: The embeddings approach shows the great potential of LLMs for enhancing protein sequence annotation, especially in viral genomics. These findings present a promising avenue for more efficient and accurate protein function inference in molecular biology.

Applications of 3D Bioprinting Technology to Brain Cells and Brain Tumor Models: Special Emphasis to Glioblastoma.

Primary brain tumor is one of the most fatal diseases. The most malignant type among them, glioblastoma (GBM), has low survival rates. Standard treatments reduce the life quality of patients due to serious side effects. Tumor aggressiveness and the unique structure of the brain render the removal of tumors and the development of new therapies challenging. To elucidate the characteristics of brain tumors and examine their response to drugs, realistic systems that mimic the tumor environment and cellular crosstalk are desperately needed. In the past decade, 3D GBM models have been presented as excellent platforms as they allowed the investigation of the phenotypes of GBM and testing innovative therapeutic strategies. In that scope, 3D bioprinting technology offers utilities such as fabricating realistic 3D bioprinted structures in a layer-by-layer manner and precisely controlled deposition of materials and cells, and they can be integrated with other technologies like the microfluidics approach. This Review covers studies that investigated 3D bioprinted brain tumor models, especially GBM using 3D bioprinting techniques and essential parameters that affect the result and quality of the study like frequently used cells, the type and physical characteristics of hydrogel, bioprinting conditions, cross-linking methods, and characterization techniques.

Assessing reproductive performance and predictive models for litter size in Landrace sows under tropical conditions.

Objective: Litter size and piglet loss at birth significantly impact piglet production and are closely associated with sow parity. Understanding how these traits vary across different parities is crucial for effective herd management. This study investigates the patterns of the number of born alive piglets (NBA), number of piglet losses (NPL), and the proportion of piglet losses (PPL) at birth in Landrace sows under tropical conditions. Additionally, it aims to identify the most suitable model for describing these patterns. Methods: A dataset comprising 2,322 consecutive reproductive records from 258 Landrace sows, spanning parities from 1 to 9, was analyzed. Modeling approaches including 2nd and 3rd degree polynomial models, the Wood gamma function, and a longitudinal model were applied at the individual level to predict NBA, NPL, and PPL. The choice of the best-fitting model was determined based on the lowest mean and standard deviation of the difference between predicted and actual values, Akaike information criterion (AIC), and Bayesian information criterion (BIC). Results: Sow parity significantly influenced NBA, NPL, and PPL (p<0.0001). NBA increased until the 4th parity and then declined. In contrast, NPL and PPL decreased until the 2nd parity and then steadily increased until the 8th parity. The 2nd and 3rd degree polynomials, and longitudinal models showed no significant differences in predicting NBA, NPL, and PPL (p>0.05). The 3rd degree polynomial model had the lowest prediction standard deviation and yielded the smallest AIC and BIC. Conclusion: The 3rd degree polynomial model offers the most suitable description of NBA, NPL, and PPL patterns. It holds promise for applications in genetic evaluations to enhance litter size and reduce piglet loss at birth in sows. These findings highlight the importance of accounting for sow parity effects in swine breeding programs, particularly in tropical conditions, to optimize piglet production and sow performance.

Prevalence and Determinants of Low Birth Weight among the Newborns in Dadra and Nagar Haveli: A Community-based Study.

Background: The world is not on track to meet the World Health Assembly (WHA) global target on Low Birth Weight (LBW). To estimate the prevalence and to identify the associated determinants of LBW among the newborns. Material and Methods: We conducted a cross-sectional study among the 364 mothers registered under the all government health facilities of Dadra & Nagar Haveli (DNH) during November 2021 to January 2022. Results: The prevalence of LBW was found to be 39%. On uni-variable logistic regression, live in relationship, caste, weight of mother, Body Mass Index (BMI), weight gain <5 kg in 2nd and 3rd trimester, high-risk pregnancy, complication present in previous pregnancy and preterm delivery, while on multi-variable logistic regression, weight gain <5 kg in 2nd and 3rd trimester (AOR 2, 95% CI 1.007-4.2) and having high-risk pregnancy (AOR 2, 95% CI 1.1-3.0) were found to be the significant predictors of LBW among the newborns. Conclusions: We conclude from the study that the prevalence of low birth weight among the newborn was high. There is a need to address maternal and child health issues like low birth weight, malnutrition and high-risk pregnancy under the RMNCAH+N program through various effective interventions. Future research should evaluate the feasibility of collaborative activities between RMNCAH+N program and the UNICEF in India.

Identifying optimal candidates for primary tumor surgery in patients with metastatic head and neck cancer.

Background: Primary tumor surgery (PTS) may enhance survival among part of patients with metastatic head and neck cancer (mHNC). Herein, a predictive model was needed to construct to identify who can gain benefit remarkably from tumor resection. Methods: Data of patients with mHNC were extracted from the Surveillance, Epidemiology, and End Results (SEER) database. The best cut-off value of age were analyzed using the X-tile software. One-to-one PSM, Kaplan-Meier method, and log-rank test were performed for survival analysis.The independent factors determined using the multivariate Cox proportional hazard regression were used to construct the nomogram. Results: A total of 1,614 patients diagnosed with mHNC were included; among them, 356 (22.0%) underwent a surgical procedure for the excision of the primary tumor. cancer-specific survival (CSS) was remarkably prolonged in the PTS group relative to the non-PTS group following PSM [Median:19 months vs. 9 months; hazard ratio (HR) 0.52, P < 0.001]. Patients with mHNC who were younger than 52 years old, had well-differentiated tumors, had T1 and N0 stages, and were married at the time of the study may have significantly benefited from PTS. In addition, we constructed a nomogram based on the factors that independently affect the CSS in multivariate Cox analysis. The nomogram showed excellent discrimination in both the training and validation sets (AUC: 0.732 and 0.738, respectively). Conclusion: A practical predictive model was constructed to determine the appropriate patients with mHNC, who would benefit from surgical resection.

Mind your prevalence!

Multiple metrics are used when assessing and validating the performance of quantitative structure-activity relationship (QSAR) models. In the case of binary classification, balanced accuracy is a metric to assess the global performance of such models. In contrast to accuracy, balanced accuracy does not depend on the respective prevalence of the two categories in the test set that is used to validate a QSAR classifier. As such, balanced accuracy is used to overcome the effect of imbalanced test sets on the model's perceived accuracy. Matthews' correlation coefficient (MCC), an alternative global performance metric, is also known to mitigate the imbalance of the test set. However, in contrast to the balanced accuracy, MCC remains dependent on the respective prevalence of the predicted categories. For simplicity, the rest of this work is based on the positive prevalence. The MCC value may be underestimated at high or extremely low positive prevalence. It contributes to more challenging comparisons between experiments using test sets with different positive prevalences and may lead to incorrect interpretations. The concept of balanced metrics beyond balanced accuracy is, to the best of our knowledge, not yet described in the cheminformatic literature. Therefore, after describing the relevant literature, this manuscript will first formally define a confusion matrix, sensitivity and specificity and then present, with synthetic data, the danger of comparing performance metrics under nonconstant prevalence. Second, it will demonstrate that balanced accuracy is the performance metric accuracy calibrated to a test set with a positive prevalence of 50% (i.e., balanced test set). This concept of balanced accuracy will then be extended to the MCC after showing its dependency on the positive prevalence. Applying the same concept to any other performance metric and widening it to the concept of calibrated metrics will then be briefly discussed. We will show that, like balanced accuracy, any balanced performance metric may be expressed as a function of the well-known values of sensitivity and specificity. Finally, a tale of two MCCs will exemplify the use of this concept of balanced MCC versus MCC with four use cases using synthetic data. SCIENTIFIC CONTRIBUTION: This work provides a formal, unified framework for understanding prevalence dependence in model validation metrics, deriving balanced metric expressions beyond balanced accuracy, and demonstrating their practical utility for common use cases. In contrast to prior literature, it introduces the derived confusion matrix to express metrics as functions of sensitivity, specificity and prevalence without needing additional coefficients. The manuscript extends the concept of balanced metrics to Matthews' correlation coefficient and other widely used performance indicators, enabling robust comparisons under prevalence shifts.

Powerful and accurate detection of temporal gene expression patterns from multi-sample multi-stage single-cell transcriptomics data with TDEseq.

We present a non-parametric statistical method called TDEseq that takes full advantage of smoothing splines basis functions to account for the dependence of multiple time points in scRNA-seq studies, and uses hierarchical structure linear additive mixed models to model the correlated cells within an individual. As a result, TDEseq demonstrates powerful performance in identifying four potential temporal expression patterns within a specific cell type. Extensive simulation studies and the analysis of four published scRNA-seq datasets show that TDEseq can produce well-calibrated p-values and up to 20% power gain over the existing methods for detecting temporal gene expression patterns.