Artificial intelligence in plant breeding.

Harnessing cutting-edge technologies to enhance crop productivity is a pivotal goal in modern plant breeding. Artificial intelligence (AI) is renowned for its prowess in big data analysis and pattern recognition, and is revolutionizing numerous scientific domains including plant breeding. We explore the wider potential of AI tools in various facets of breeding, including data collection, unlocking genetic diversity within genebanks, and bridging the genotype-phenotype gap to facilitate crop breeding. This will enable the development of crop cultivars tailored to the projected future environments. Moreover, AI tools also hold promise for refining crop traits by improving the precision of gene-editing systems and predicting the potential effects of gene variants on plant phenotypes. Leveraging AI-enabled precision breeding can augment the efficiency of breeding programs and holds promise for optimizing cropping systems at the grassroots level. This entails identifying optimal inter-cropping and crop-rotation models to enhance agricultural sustainability and productivity in the field.

The NLR immune receptor ADR1 and lipase-like proteins EDS1 and PAD4 mediate stomatal immunity in Nicotiana benthamiana and Arabidopsis.

In the presence of pathogenic bacteria, plants close their stomata to prevent pathogen entry. Intracellular nucleotide-binding leucine-rich repeat (NLR) immune receptors recognize pathogenic effectors and activate effector-triggered immune responses. However, the regulatory and molecular mechanisms of stomatal immunity involving NLR immune receptors are unknown. Here, we show that the Nicotiana benthamiana RPW8-NLR central immune receptor ACTIVATED DISEASE RESISTANCE 1 (NbADR1), together with the key immune proteins ENHANCED DISEASE SUSCEPTIBILITY 1 (NbEDS1) and PHYTOALEXIN DEFICIENT 4 (NbPAD4), plays an essential role in bacterial pathogen- and flg22-induced stomatal immunity by regulating the expression of salicylic acid (SA) and abscisic acid (ABA) biosynthesis or response-related genes. NbADR1 recruits NbEDS1 and NbPAD4 in stomata to form a stomatal immune response complex. The transcription factor NbWRKY40e, in association with NbEDS1 and NbPAD4, modulates the expression of SA and ABA biosynthesis or response-related genes to influence stomatal immunity. NbADR1, NbEDS1, and NbPAD4 are required for the pathogen infection-enhanced binding of NbWRKY40e to the ISOCHORISMATE SYNTHASE 1 promoter. Moreover, the ADR1-EDS1-PAD4 module regulates stomatal immunity in Arabidopsis (Arabidopsis thaliana). Collectively, our findings show the pivotal role of the core intracellular immune receptor module ADR1-EDS1-PAD4 in stomatal immunity, which enables plants to limit pathogen entry.

Fractional antiferromagnetic skyrmion lattice induced by anisotropic couplings.

Magnetic skyrmions are topological solitons with a nanoscale winding spin texture that hold promise for spintronics applications1-4. Skyrmions have so far been observed in a variety of magnets that exhibit nearly parallel alignment for neighbouring spins, but theoretically skyrmions with anti-parallel neighbouring spins are also possible. Such antiferromagnetic skyrmions may allow more flexible control than conventional ferromagnetic skyrmions5-10. Here, by combining neutron scattering measurements and Monte Carlo simulations, we show that a fractional antiferromagnetic skyrmion lattice is stabilized in MnSc2S4 through anisotropic couplings. The observed lattice is composed of three antiferromagnetically coupled sublattices, and each sublattice is a triangular skyrmion lattice that is fractionalized into two parts with an incipient meron (half-skyrmion) character11,12. Our work demonstrates that the theoretically proposed antiferromagnetic skyrmions can be stabilized in real materials and represents an important step towards their implementation in spintronic devices.

m6Aexpress-enet: Predicting the regulatory expression m6A sites by an enet-regularization negative binomial regression model.

As the most abundant mRNA modification, m6A controls and influences many aspects of mRNA metabolism including the mRNA stability and degradation. However, the role of specific m6A sites in regulating gene expression still remains unclear. In additional, the multicollinearity problem caused by the correlation of methylation level of multiple m6A sites in each gene could influence the prediction performance. To address the above challenges, we propose an elastic-net regularized negative binomial regression model (called m6Aexpress-enet) to predict which m6A site could potentially regulate its gene expression. Comprehensive evaluations on simulated datasets demonstrate that m6Aexpress-enet could achieve the top prediction performance. Applying m6Aexpress-enet on real MeRIP-seq data from human lymphoblastoid cell lines, we have uncovered the complex regulatory pattern of predicted m6A sites and their unique enrichment pathway of the constructed co-methylation modules. m6Aexpress-enet proves itself as a powerful tool to enable biologists to discover the mechanism of m6A regulatory gene expression. Furthermore, the source code and the step-by-step implementation of m6Aexpress-enet is freely accessed at https://github.com/tengzhangs/m6Aexpress-enet.

Identification of STAT3 as a biomarker for cellular senescence in liver fibrosis: A bioinformatics and experimental validation study.

BACKGROUND: Cellular senescence is associated with a dysregulated inflammatory response, which is an important driver of the development of liver fibrosis (LF). This study aimed to investigate the effect of cellular senescence on LF and identify potential key biomarkers through bioinformatics analysis combined with validation experiments in vivo and in vitro. METHODS: The Gene Expression Omnibus (GEO) database and GeneCards database were used to download the LF dataset and the aging-related gene set, respectively. Functional enrichment analysis of differential genes was then performed using GO and KEGG. Hub genes were further screened using Cytoscape's cytoHubba. Diagnostic values for hub genes were evaluated with a receiver operating characteristic (ROC) curve. Next, CIBERSORTx was used to estimate immune cell types and ratios. Finally, in vivo and in vitro experiments validated the results of the bioinformatics analysis. Moreover, molecular docking was used to simulate drug-gene interactions. RESULTS: A total of 44 aging-related differentially expressed genes (AgDEGs) were identified, and enrichment analysis showed that these genes were mainly enriched in inflammatory and immune responses. PPI network analysis identified 6 hub AgDEGs (STAT3, TNF, MMP9, CD44, TGFB1, and TIMP1), and ROC analysis showed that they all have good diagnostic value. Immune infiltration suggested that hub AgDEGs were significantly associated with M1 macrophages or other immune cells. Notably, STAT3 was positively correlated with α-SMA, COL1A1, IL-6 and IL-1ß, and was mainly expressed in hepatocytes (HCs). Validation experiments showed that STAT3 expression was upregulated and cellular senescence was increased in LF mice. A co-culture system of HCs and hepatic stellate cells (HSCs) further revealed that inhibiting STAT3 reduced HCs senescence and suppressed HSCs activation. In addition, molecular docking revealed that STAT3 was a potential drug therapy target. CONCLUSIONS: STAT3 may be involved in HCs senescence and promote HSCs activation, which in turn leads to the development of LF. Our findings suggest that STAT3 could be a potential biomarker for LF.

Deep learning-enabled discovery and characterization of HKT genes in Spartina alterniflora.

Spartina alterniflora is a halophyte that can survive in high-salinity environments, and it is phylogenetically close to important cereal crops, such as maize and rice. It is of scientific interest to understand why S. alterniflora can live under such extremely stressful conditions. The molecular mechanism underlying its high-saline tolerance is still largely unknown. Here we investigated the possibility that high-affinity K+ transporters (HKTs), which function in salt tolerance and maintenance of ion homeostasis in plants, are responsible for salt tolerance in S. alterniflora. To overcome the imprecision and unstable of the gene screening method caused by the conventional sequence alignment, we used a deep learning method, DeepGOPlus, to automatically extract sequence and protein characteristics from our newly assemble S. alterniflora genome to identify SaHKTs. Results showed that a total of 16 HKT genes were identified. The number of S. alterniflora HKTs (SaHKTs) is larger than that in all other investigated plant species except wheat. Phylogenetically related SaHKT members had similar gene structures, conserved protein domains and cis-elements. Expression profiling showed that most SaHKT genes are expressed in specific tissues and are differentially expressed under salt stress. Yeast complementation expression analysis showed that type I members SaHKT1;2, SaHKT1;3 and SaHKT1;8 and type II members SaHKT2;1, SaHKT2;3 and SaHKT2;4 had low-affinity K+ uptake ability and that type II members showed stronger K+ affinity than rice and Arabidopsis HKTs, as well as most SaHKTs showed preference for Na+ transport. We believe the deep learning-based methods are powerful approaches to uncovering new functional genes, and the SaHKT genes identified are important resources for breeding new varieties of salt-tolerant crops.

SMEK1 ablation promotes glucose uptake and improves obesity-related metabolic dysfunction via AMPK signaling pathway.

Obesity has become a major risk of global public health. SMEK1 is also known as a regulatory subunit of protein phosphatase 4 (PP4). Both PP4 and SMEK1 have been clarified in many metabolic functions, including the regulation of hepatic gluconeogenesis and glucose transporter gene expression in yeast. Whether SMEK1 participates in obesity and the broader metabolic role in mammals is unknown. Thus, we investigated the function of SMEK1 in white adipose tissue and glucose uptake. GWAS/GEPIA/GEO database was used to analyze the correlation between SMEK1 and metabolic phenotypes/lipid metabolism-related genes/obesity. Smek1 KO mice were generated to identify the role of SMEK1 in obesity and glucose homeostasis. Cell culture and differentiation of stromal-vascular fractions (SVFs) and 3T3-L1 were used to determine the mechanism. 2-NBDG was used to measure the glucose uptake. Compound C was used to confirm the role of AMPK. We elucidated that SMEK1 was correlated with obesity and adipogenesis. Smek1 deletion enhanced adipogenesis in both SVFs and 3T3-L1. Smek1 KO protected mice from obesity and had protective effects on metabolic disorders, including insulin resistance and inflammation. Smek1 KO mice had lower levels of fasting serum glucose. We found that SMEK1 ablation promoted glucose uptake by increasing p-AMPKα(T172) and the transcription of Glut4 when the effect on AMPK-regulated glucose uptake was due to the PP4 catalytic subunits (PPP4C). Our findings reveal a novel role of SMEK1 in obesity and glucose homeostasis, providing a potential new therapeutic target for obesity and metabolic dysfunction.NEW & NOTEWORTHY Our study clarified the relationship between SMEK1 and obesity for the first time and validated the conclusion in multiple ways by combining available data from public databases, human samples, and animal models. In addition, we clarified the role of SMEK1 in glucose uptake, providing an in-depth interpretation for the study of its function in glucose metabolism.

Plexin D1 negatively regulates macrophage-derived foam cell migration via the focal adhesion kinase/Paxillin pathway.

BACKGROUND: Macrophage-derived foam cell formation is a hallmark of atherosclerosis and is retained during plaque formation. Strategies to inhibit the accumulation of these cells hold promise as viable options for treating atherosclerosis. Plexin D1 (PLXND1), a member of the Plexin family, has elevated expression in atherosclerotic plaques and correlates with cell migration; however, its role in macrophages remains unclear. We hypothesize that the guidance receptor PLXND1 negatively regulating macrophage mobility to promote the progression of atherosclerosis. METHODS: We utilized a mouse model of atherosclerosis based on a high-fat diet and an ox-LDL- induced foam cell model to assess PLXND1 levels and their impact on cell migration. Through western blotting, Transwell assays, and immunofluorescence staining, we explored the potential mechanism by which PLXND1 mediates foam cell motility in atherosclerosis. RESULTS: Our study identifies a critical role for PLXND1 in atherosclerosis plaques and in a low-migration capacity foam cell model induced by ox-LDL. In the aortic sinus plaques of ApoE-/- mice, immunofluorescence staining revealed significant upregulation of PLXND1 and Sema3E, with colocalization in macrophages. In macrophages treated with ox-LDL, increased expression of PLXND1 led to reduced pseudopodia formation and decreased migratory capacity. PLXND1 is involved in regulating macrophage migration by modulating the phosphorylation levels of FAK/Paxillin and downstream CDC42/PAK. Additionally, FAK inhibitors counteract the ox-LDL-induced migration suppression by modulating the phosphorylation states of FAK, Paxillin and their downstream effectors CDC42 and PAK. CONCLUSION: Our findings indicate that PLXND1 plays a role in regulating macrophage migration by modulating the phosphorylation levels of FAK/Paxillin and downstream CDC42/PAK to promoting atherosclerosis.

A Bayesian inference transcription factor activity model for the analysis of single-cell transcriptomes.

Single-cell RNA sequencing (scRNA-seq) has emerged as a powerful experimental approach to study cellular heterogeneity. One of the challenges in scRNA-seq data analysis is integrating different types of biological data to consistently recognize discrete biological functions and regulatory mechanisms of cells, such as transcription factor activities and gene regulatory networks in distinct cell populations. We have developed an approach to infer transcription factor activities from scRNA-seq data that leverages existing biological data on transcription factor binding sites. The Bayesian inference transcription factor activity model (BITFAM) integrates ChIP-seq transcription factor binding information into scRNA-seq data analysis. We show that the inferred transcription factor activities for key cell types identify regulatory transcription factors that are known to mechanistically control cell function and cell fate. The BITFAM approach not only identifies biologically meaningful transcription factor activities, but also provides valuable insights into underlying transcription factor regulatory mechanisms.

The Spartina alterniflora genome sequence provides insights into the salt-tolerance mechanisms of exo-recretohalophytes.

Spartina alterniflora is an exo-recretohalophyte Poaceae species that is able to grow well in seashore, but the genomic basis underlying its adaptation to salt tolerance remains unknown. Here, we report a high-quality, chromosome-level genome assembly of S. alterniflora constructed through PacBio HiFi sequencing, combined with high-throughput chromosome conformation capture (Hi-C) technology and Illumina-based transcriptomic analyses. The final 1.58 Gb genome assembly has a contig N50 size of 46.74 Mb. Phylogenetic analysis suggests that S. alterniflora diverged from Zoysia japonica approximately 21.72 million years ago (MYA). Moreover, whole-genome duplication (WGD) events in S. alterniflora appear to have expanded gene families and transcription factors relevant to salt tolerance and adaptation to saline environments. Comparative genomics analyses identified numerous species-specific genes, significantly expanded genes and positively selected genes that are enriched for 'ion transport' and 'response to salt stress'. RNA-seq analysis identified several ion transporter genes including the high-affinity K+ transporters (HKTs), SaHKT1;2, SaHKT1;3 and SaHKT1;8, and high copy number of Salt Overly Sensitive (SOS) up-regulated under high salt conditions, and the overexpression of SaHKT2;4 in Arabidopsis thaliana conferred salt tolerance to the plant, suggesting specialized roles for S. alterniflora to adapt to saline environments. Integrated metabolomics and transcriptomics analyses revealed that salt stress activate glutathione metabolism, with differential expressions of several genes such as γ-ECS, GSH-S, GPX, GST and PCS in the glutathione metabolism. This study suggests several adaptive mechanisms that could contribute our understanding of evolutional basis of the halophyte.

Association between premature vascular smooth muscle cells senescence and vascular inflammation in Takayasu's arteritis.

OBJECTIVES: Arterial wall inflammation and remodelling are the characteristic features of Takayasu's arteritis (TAK). It has been proposed that vascular smooth muscle cells (VSMCs) are the main targeted cells of inflammatory damage and participate in arterial remodelling in TAK. Whether VSMCs are actively involved in arterial wall inflammation has not been elucidated. Studies have shown that cellular senescence in tissue is closely related to local inflammation persistence. We aimed to investigate whether VSMCs senescence contributes to vascular inflammation and the prosenescent factors in TAK. METHODS: VSMCs senescence and senescence-associated secretory phenotype were detected by histological examination, bulk RNA-Seq and single-cell RNA-seq conducted on vascular surgery samples of TAK patients. The key prosenescent factors and the downstream signalling pathway were investigated in a series of in vitro and ex vivo experiments. RESULTS: Histological findings, primary cell culture and transcriptomic analyses demonstrated that VSMCs of TAK patients had the features of premature senescence and contributed substantially to vascular inflammation by upregulating the expression of senescence-associated inflammatory cytokines. IL-6 was found to be the critical cytokine that drove VSMCs senescence and senescence-associated mitochondrial dysfunction in TAK. Mechanistically, IL-6-induced non-canonical mitochondrial localisation of phosphorylated STAT3 (Tyr705) prevented mitofusin 2 (MFN2) from proteasomal degradation, and subsequently promoted senescence-associated mitochondrial dysfunction and VSMCs senescence. Mitochondrial STAT3 or MFN2 inhibition ameliorated VSMCs senescence in ex vivo cultured arteries of TAK patients. CONCLUSIONS: VSMCs present features of cellular senescence and are actively involved in vascular inflammation in TAK. Vascular IL-6-mitochondrial STAT3-MFN2 signalling is an important driver of VSMCs senescence.

Identification of a novel and plant height-independent QTL for coleoptile length in barley and validation of its effect using near isogenic lines.

KEY MESSAGE: This study reported the identification and validation of novel QTL conferring coleoptile length in barley and predicted candidate genes underlying the largest effect QTL based on orthologous analysis and comparison of the whole genome assemblies for both parental genotypes of the mapping population. Coleoptile length (CL) is one of the most important agronomic traits in cereal crops due to its direct influence on the optimal depth for seed sowing which facilitates better seedling establishment. Varieties with longer coleoptiles are preferred in drought-prone areas where less moisture maintains at the top layer of the soil. Compared to wheat, genetic study on coleoptile length is limited in barley. Here, we reported a study on detecting the genomic regions associated with CL in barley by assessing a population consisting of 201 recombinant inbred lines. Four putative QTL conferring CL were consistently identified on chromosomes 1H, 5H, 6H, and 7H in each of the trials conducted. Of these QTL, the two located on chromosomes 5H and 6H (designated as Qcl.caf-5H and Qcl.caf-6H) are likely novel and Qcl.caf-5H showed the most significant effect explaining up to 30.9% of phenotypic variance with a LOD value of 15.1. To further validate the effect of this putative QTL, five pairs of near isogenic lines (NILs) were then developed and assessed. Analysis of the NILs showed an average difference of 21.0% in CL between the two isolines. Notably, none of the other assessed morphological characteristics showed consistent differences between the two isolines for each pair of the NILs. Candidate genes underlying the Qcl.caf-5H locus were also predicted by employing orthologous analysis and comparing the genome assemblies for both parental genotypes of the mapping population in the present study. Taken together, these findings expand our understanding on genetic basis of CL and will be indicative for further gene cloning and functional analysis underly this locus in barley.

Transcriptomic insights into shared responses to Fusarium crown rot infection and drought stresses in bread wheat (Triticum aestivum L.).

KEY MESSAGE: Shared changes in transcriptomes caused by Fusarium crown rot infection and drought stress were investigated based on a single pair of near-isogenic lines developed for a major locus conferring tolerance to both stresses. Fusarium crown rot (FCR) is a devastating disease in many areas of cereal production worldwide. It is well-known that drought stress enhances FCR severity but possible molecular relationship between these two stresses remains unclear. To investigate their relationships, we generated several pairs of near isogenic lines (NILs) targeting a locus conferring FCR resistance on chromosome 2D in bread wheat. One pair of these NILs showing significant differences between the two isolines for both FCR resistance and drought tolerance was used to investigate transcriptomic changes in responsive to these two stresses. Our results showed that the two isolines likely deployed different strategies in dealing with the stresses, and significant differences in expressed gene networks exist between the two time points of drought stresses evaluated in this study. Nevertheless, results from analysing Gene Ontology terms and transcription factors revealed that similar regulatory frameworks were activated in coping with these two stresses. Based on the position of the targeted locus, changes in expression following FCR infection and drought stresses, and the presence of non-synonymous variants between the two isolines, several candidate genes conferring resistance or tolerance to these two types of stresses were identified. The NILs generated, the large number of DEGs with single-nucleotide polymorphisms detected between the two isolines, and the candidate genes identified would be invaluable in fine mapping and cloning the gene(s) underlying the targeted locus.

Quantitative turbidimetric characterization of stabilized complex coacervate dispersions.

Stabilizing complex coacervate microdroplets is desirable due to their various applications, such as bioreactors, drug delivery vehicles, and encapsulants. Here, we present quantitative characterization of complex coacervate dispersion stability inferred by turbidimetry measurements. The stability of the dispersions is shown to be modulated by the concentrations of comb polyelectrolyte (cPE) stabilizers and salt. We demonstrate cPEs as effective stabilizers for complex coacervate dispersions independent of the chemistry or length of the constituent polyelectrolytes, salts, or preparation routes. By monitoring the temporal evolution of dispersion turbidity, we show that cPEs suppress microdroplet coalescence with minimal change in microdroplet sizes over 48 hours, even at salt concentrations up to 300 mM. The number density and average microdroplet size are shown to be controlled by varying the cPE and salt concentrations. Lastly, turbidity maps, akin to binodal phase maps, depict an expansion of the turbid two-phase region and an increase in the salt resistance of the coacervates upon the introduction of cPEs. The coacervate salt resistance is shown to increase by >3×, and this increase is maintained for up to 15 days, demonstrating that cPEs impart higher salt resistance over extended durations.

Foam film stratification, viscosity, and small-angle X-ray scattering of micellar SDS solutions over an extended concentration range (1< c/CMC < 75).

Ultrathin foam films (thickness, h < 100 nm) containing micelles undergo drainage via stratification manifested as coexisting thick-thin flat regions, nanoscopic non-flat topography, and the stepwise decrease in film thickness that yields a characteristic step-size. Most studies characterize the variation in step size and stratification kinetics in micellar foam films in a limited concentration range, c/CMC < 12.5 (c < 100 mM). Likewise, most scattering studies characterize micelle dimensions, intermicellar distance, and volume fraction in bulk aqueous SDS solutions in this limited concentration range. In this contribution, we show drainage via stratification can be observed for concentrations up to c/CMC < 75 (c < 600 mM). Understanding the stratification behavior of freely draining micellar films with sodium dodecyl sulfate (SDS) concentration varying in the range 10 mM ≤ cSDS ≤ 600 mM is essential for molecular engineering, consumer product formulations, and controlling foaming in industrial processes. Here, we visualize and analyze nanoscopic thickness variations and transitions in stratifying foam films using Interferometry Digital Imaging Optical Microscopy (IDIOM) protocols. We compare step size obtained from foam stratification to micelle dimension, micelle volume fraction, and intermicellar distance obtained from small angle X-ray scattering studies. Even though the volume fraction increases and approaches 25% at c = 600 mM, the solution viscosity only increases by a factor of four compared to the solvent, consistent with the findings from both stratification and scattering studies. These comparisons allow us to explore the effect of micelle size, morphology, and intermicellar interactions on supramolecular oscillatory structural disjoining pressure, which influences the stratification behavior of draining foam films containing micelles under confinement.

Impacts of stream drying depend on stream network size and location of drying.

Stream drying is increasing globally, with widespread impacts on stream ecosystems. Here, we investigated how the impacts of drying on stream ecosystem connectivity might depend on stream network size and the location of drying within the stream network. Using 11 stream networks from across the United States, we simulated drying scenarios in which we varied the location and spatial extent of drying. We found that the rate of connectivity loss varied with stream network size, such that larger stream networks lost connectivity more rapidly than smaller stream networks. We also found that the rate of connectivity loss varied with the location of drying. When drying occurred in the mainstem, even small amounts of drying resulted in rapid losses in ecosystem connectivity. When drying occurred in headwater reaches, small amounts of drying had little impact on connectivity. Beyond a certain threshold, however, connectivity declined rapidly with further increases in drying. Given the increasing stream drying worldwide, our findings underscore the need for managers to be particularly vigilant about fragmentation when managing at large spatial scales and when stream drying occurs in mainstem reaches.

Atroposelective Synthesis of Axially Chiral Diaryl Ethers by N-Heterocyclic-Carbene-Catalyzed Sequentially Desymmetric/Kinetic Resolution Process.

We describe herein an N-heterocyclic-carbene-catalyzed atroposelective synthesis of axially chiral diaryl ethers. Through a sequentially enantioselective desymmetric process and a kinetic resolution process, the products could be constructed in good yields with excellent enantiopurities. Both alcohols and phenols were compatible with this catalytic system. The axially chiral carboxylic acids derived from the esters were proven to be potential chiral ligands for asymmetric synthesis, for example, Rh(III)-catalyzed enantioselective C-H functionalization.

GPT-4 passes the bar exam.

In this paper, we experimentally evaluate the zero-shot performance of GPT-4 against prior generations of GPT on the entire uniform bar examination (UBE), including not only the multiple-choice multistate bar examination (MBE), but also the open-ended multistate essay exam (MEE) and multistate performance test (MPT) components. On the MBE, GPT-4 significantly outperforms both human test-takers and prior models, demonstrating a 26% increase over ChatGPT and beating humans in five of seven subject areas. On the MEE and MPT, which have not previously been evaluated by scholars, GPT-4 scores an average of 4.2/6.0 when compared with much lower scores for ChatGPT. Graded across the UBE components, in the manner in which a human test-taker would be, GPT-4 scores approximately 297 points, significantly in excess of the passing threshold for all UBE jurisdictions. These findings document not just the rapid and remarkable advance of large language model performance generally, but also the potential for such models to support the delivery of legal services in society. This article is part of the theme issue 'A complexity science approach to law and governance'.

Deep learning uncertainty quantification for clinical text classification.

INTRODUCTION: Machine learning algorithms are expected to work side-by-side with humans in decision-making pipelines. Thus, the ability of classifiers to make reliable decisions is of paramount importance. Deep neural networks (DNNs) represent the state-of-the-art models to address real-world classification. Although the strength of activation in DNNs is often correlated with the network's confidence, in-depth analyses are needed to establish whether they are well calibrated. METHOD: In this paper, we demonstrate the use of DNN-based classification tools to benefit cancer registries by automating information extraction of disease at diagnosis and at surgery from electronic text pathology reports from the US National Cancer Institute (NCI) Surveillance, Epidemiology, and End Results (SEER) population-based cancer registries. In particular, we introduce multiple methods for selective classification to achieve a target level of accuracy on multiple classification tasks while minimizing the rejection amount-that is, the number of electronic pathology reports for which the model's predictions are unreliable. We evaluate the proposed methods by comparing our approach with the current in-house deep learning-based abstaining classifier. RESULTS: Overall, all the proposed selective classification methods effectively allow for achieving the targeted level of accuracy or higher in a trade-off analysis aimed to minimize the rejection rate. On in-distribution validation and holdout test data, with all the proposed methods, we achieve on all tasks the required target level of accuracy with a lower rejection rate than the deep abstaining classifier (DAC). Interpreting the results for the out-of-distribution test data is more complex; nevertheless, in this case as well, the rejection rate from the best among the proposed methods achieving 97% accuracy or higher is lower than the rejection rate based on the DAC. CONCLUSIONS: We show that although both approaches can flag those samples that should be manually reviewed and labeled by human annotators, the newly proposed methods retain a larger fraction and do so without retraining-thus offering a reduced computational cost compared with the in-house deep learning-based abstaining classifier.

Foam film stratification studies probe intermicellar interactions.

Ultrathin foam films containing supramolecular structures like micelles in bulk and adsorbed surfactant at the liquid-air interface undergo drainage via stratification. At a fixed surfactant concentration, the stepwise decrease in the average film thickness of a stratifying micellar film yields a characteristic step size that also describes the quantized thickness difference between coexisting thick-thin flat regions. Even though many published studies claim that step size equals intermicellar distance obtained using scattering from bulk solutions, we found no reports of a direct comparison between the two length scales. It is well established that step size is inversely proportional to the cubic root of surfactant concentration but cannot be estimated by adding micelle size to Debye length, as the latter is inversely proportional to the square root of surfactant concentration. In this contribution, we contrast the step size obtained from analysis of nanoscopic thickness variations and transitions in stratifying foam films using Interferometry Digital Imaging Optical Microscopy (IDIOM) protocols, that we developed, with the intermicellar distance obtained using small-angle X-ray scattering. We find that stratification driven by the confinement-induced layering of micelles within the liquid-air interfaces of a foam film provides a sensitive probe of non-DLVO (Derjaguin-Landau-Verwey-Overbeek) supramolecular oscillatory structural forces and micellar interactions.