Explore the vast range of titles available.

Breast cancer remains a significant global health challenge among women, with histopathological image analysis playing a critical role in early detection. However, existing diagnostic models often struggle to extract complex patterns from high-resolution tissue images, limiting their diagnostic accuracy and generalization. This study aims to develop a high-performance deep learning framework for accurate classification of breast cancer in histopathology images, addressing limitations in feature extraction and spatial dependency modeling. A hybrid deep learning model is proposed, integrating EfficientNetV2 for multi-scale feature extraction with a Gated Recurrent Unit (GRU) enhanced by an attention mechanism to model sequential dependencies. The model is trained and evaluated using the BreakHisand Camelyon17 dataset at 200× magnification. Evaluation metrics include precision, recall, F1-score, specificity, Intersection over Union (IoU), and accuracy. The proposed model achieved superior performance compared to existing architectures such as AlexNet, DenseNet, MobileNetV3, and EfficientNet. It attained a precision of 98.15%, recall of 95.68%, F1-score of 96.82%, specificity of 96%, IoU of 93.99%, and accuracy of 95.72% on the test set. The integration of EfficientNetV2 with GRU and attention mechanisms enables effective learning of spatial and contextual features, enhancing the accuracy and interpretability of breast cancer classification from histopathology images. This framework shows strong potential for aiding pathologists in real-time diagnostic workflows.

Increasing evidence suggests that chromatin modifications have important roles in modulating constitutive or alternative splicing. Here we demonstrate that the PWWP domain of the chromatin-associated protein Psip1/Ledgf can specifically recognize tri-methylated H3K36 and that, like this histone modification, the Psip1 short (p52) isoform is enriched at active genes. We show that the p52, but not the long (p75), isoform of Psip1 co-localizes and interacts with Srsf1 and other proteins involved in mRNA processing. The level of H3K36me3 associated Srsf1 is reduced in Psip1 mutant cells and alternative splicing of specific genes is affected. Moreover, we show altered Srsf1 distribution around the alternatively spliced exons of these genes in Psip1 null cells. We propose that Psip1/p52, through its binding to both chromatin and splicing factors, might act to modulate splicing.

Wendy Bickmore, Madapura Pradeepa and colleagues identify a new class of active enhancers marked by histones with modifications on residues in the globular domain. They find that H3K64ac and H3K122ac are markers for active promoters and enhancers in embryonic stem cells and human cancer cell lines. Histone acetylation is generally associated with active chromatin, but most studies have focused on the acetylation of histone tails. Various histone H3 and H4 tail acetylations mark the promoters of active genes 1 . These modifications include acetylation of histone H3 at lysine 27 (H3K27ac), which blocks Polycomb-mediated trimethylation of H3K27 (H3K27me3) 2 . H3K27ac is also widely used to identify active enhancers 3 , 4 , and the assumption has been that profiling H3K27ac is a comprehensive way of cataloguing the set of active enhancers in mammalian cell types. Here we show that acetylation of lysine residues in the globular domain of histone H3 (lysine 64 (H3K64ac) and lysine 122 (H3K122ac)) marks active gene promoters and also a subset of active enhancers. Moreover, we find a new class of active functional enhancers that is marked by H3K122ac but lacks H3K27ac. This work suggests that, to identify enhancers, a more comprehensive analysis of histone acetylation is required than has previously been considered.

R-loops that accumulate at transcription sites pose a persistent threat to genome integrity. PSIP1 is a chromatin protein associated with transcriptional elongation complex, possesses histone chaperone activity, and is implicated in recruiting RNA processing and DNA repair factors to transcription sites. Here, we show that PSIP1 interacts with R-loops and other proteins involved in R-loop homeostasis, including PARP1. Genome-wide mapping of PSIP1, R-loops and γ-H2AX in PSIP1-depleted human and mouse cell lines revealed an accumulation of R-loops and DNA damage at gene promoters in the absence of PSIP1. R-loop accumulation causes local transcriptional arrest and transcription-replication conflict, leading to DNA damage. PSIP1 depletion increases 53BP1 foci and reduces RAD51 foci, suggesting altered DNA repair choice. Furthermore, PSIP1 depletion increases the sensitivity of cancer cells to PARP1 inhibitors and DNA-damaging agents that induce R-loop-induced DNA damage. These findings provide insights into the mechanism through which PSIP1 maintains genome integrity at the site of transcription. R-loop accumulation at transcription sites poses a persistent threat to genome integrity. Here the authors demonstrate a role for PSIP1/LEDGF protein in reducing R-loop levels at the site of transcription and preventing transcription replication conflict to maintain genome integrity.

Due to the handling of flammable materials, offshore oil facilities are at risk of fire, endangering crew and structure. This study aims to predict fire load and structural response, focusing on a hydrocarbon fire on an offshore platform in the Gulf of Mexico. Simulations were conducted in sections due to the platform’s size, with fire load estimated for the considered fire scenario. The fire dynamic simulator (FDS) using PyroSim, a CFD tool, was used to estimate time-temperature curves, with mesh size determined by a convergence study and wind effects considered. The FDS simulation data was input into Abaqus for a detailed structural response analysis. A heat transfer analysis simulated node temperatures during fire exposure, followed by a structural analysis considering high-temperature material properties. The structural response at critical points was observed and discussed. The discussion revolves around the fire rating derived from the study’s results.

Cornelia de Lange syndrome is a multisystem developmental disorder typically caused by mutations in the gene encoding the cohesin loader NIPBL. The associated phenotype is generally assumed to be the consequence of aberrant transcriptional regulation. Recently, we identified a missense mutation in BRD4 associated with a Cornelia de Lange-like syndrome that reduces BRD4 binding to acetylated histones. Here we show that, although this mutation reduces BRD4-occupancy at enhancers it does not affect transcription of the pluripotency network in mouse embryonic stem cells. Rather, it delays the cell cycle, increases DNA damage signalling, and perturbs regulation of DNA repair in mutant cells. This uncovers a role for BRD4 in DNA repair pathway choice. Furthermore, we find evidence of a similar increase in DNA damage signalling in cells derived from NIPBL-deficient individuals, suggesting that defective DNA damage signalling and repair is also a feature of typical Cornelia de Lange syndrome. Cornelia de Lange syndrome is a developmental disorder typically caused by mutations in the gene encoding the cohesin loader NIPBL. The authors, here, by analysing previously identified mutations in BRD4 associated with the disease, reveal that a BRD4 mutation affects DNA damage signalling, and perturbs regulation of DNA repair in mutant cells.

Chronic kidney disease (CKD) is a gradual decline in renal function that can lead to kidney damage or failure. As the disease progresses, it becomes harder to diagnose. Using routine doctor consultation data to evaluate various stages of CKD could aid in early detection and prompt intervention. To this end, researchers propose a strategy for categorizing CKD using an optimization technique inspired by the learning process. Artificial intelligence has the potential to make many things in the world seem possible, even causing surprise with its capabilities. Some doctors are looking forward to advancements in technology that can scan a patient’s body and analyse their diseases. In this regard, advanced machine learning algorithms have been developed to detect the presence of kidney disease. This research presents a novel deep learning model, which combines a fuzzy deep neural network, for the recognition and prediction of kidney disease. The results show that the proposed model has an accuracy of 99.23%, which is better than existing methods. Furthermore, the accuracy of detecting chronic disease can be confirmed without doctor involvement as future work. Compared to existing information mining classifications, the proposed approach shows improved accuracy in classification, precision, F-measure, and sensitivity metrics.

Background Ribosomal DNA (rDNA) displays substantial inter-individual genetic variation in human and mouse. A systematic analysis of how this variation impacts epigenetic states and expression of the rDNA has thus far not been performed. Results Using a combination of long- and short-read sequencing, we establish that 45S rDNA units in the C57BL/6J mouse strain exist as distinct genetic haplotypes that influence the epigenetic state and transcriptional output of any given unit. DNA methylation dynamics at these haplotypes are dichotomous and life-stage specific: at one haplotype, the DNA methylation state is sensitive to the in utero environment, but refractory to post-weaning influences, whereas other haplotypes entropically gain DNA methylation during aging only. On the other hand, individual rDNA units in human show limited evidence of genetic haplotypes, and hence little discernible correlation between genetic and epigenetic states. However, in both species, adjacent units show similar epigenetic profiles, and the overall epigenetic state at rDNA is strongly positively correlated with the total rDNA copy number. Analysis of different mouse inbred strains reveals that in some strains, such as 129S1/SvImJ, the rDNA copy number is only approximately 150 copies per diploid genome and DNA methylation levels are < 5%. Conclusions Our work demonstrates that rDNA-associated genetic variation has a considerable influence on rDNA epigenetic state and consequently rRNA expression outcomes. In the future, it will be important to consider the impact of inter-individual rDNA (epi)genetic variation on mammalian phenotypes and diseases.

Long noncoding RNAs (lncRNAs) have been implicated in various biological functions including the regulation of gene expression, however, the functionality of lncRNAs is not clearly understood and conflicting conclusions have often been reached when comparing different methods to investigate them. Moreover, little is known about the upstream regulation of lncRNAs. Here we show that the short isoform (p52) of a transcriptional co-activator-PC4 and SF2 interacting protein (Psip1), which is known to be involved in linking transcription to RNA processing, specifically regulates the expression of the lncRNA Hottip-located at the 5' end of the Hoxa locus. Using both knockdown and knockout approaches we show that Hottip expression is required for activation of the 5' Hoxa genes (Hoxa13 and Hoxa10/11) and for retaining Mll1 at the 5' end of Hoxa. Moreover, we demonstrate that artificially inducing Hottip expression is sufficient to activate the 5' Hoxa genes and that Hottip RNA binds to the 5' end of Hoxa. By engineering premature transcription termination, we show that it is the Hottip lncRNA molecule itself, not just Hottip transcription that is required to maintains active expression of posterior Hox genes. Our data show a direct role for a lncRNA molecule in regulating the expression of developmentally-regulated mRNA genes in cis.