Abstract To satisfy the capacity requirements of future mobile systems, under-utilized millimeter wave frequencies can be efficiently exploited by employing massive MIMO technology with highly directive beamforming.Hybrid analog-digital beamforming has been recognised as a promising approach for large-scale MIMO implementations with a reduced number of costly and power-hungry RF chains.In comparison to fully connected architecture, hybrid beamforming (HBF) with partially connected RF architecture is particularly appealing for the practical implementation due to less complex RF power division and combining networks.In this paper, we first formulate single- and multi-user rate maximization problems as weighted minimum mean square error (WMMSE) and derive solutions for hybrid beamformers using alternating optimization.The algorithms are designed for the full-array- and sub-array-based processing strategies of partially connected HBF architecture.In addition to the rate maximizing WMMSE solutions, we propose lower complexity sub-array-based zero-forcing algorithms.The performance of the proposed algorithms is evaluated in two different channel models, i.e., a simple geometric model and a realistic statistical millimeter wave model known as NYUSIM.The performance results of the WMMSE HBF algorithms are meant to reveal the potential of partially connected HBF and serve as upper bounds for lower complexity methods.Numerical results imply that properly designed partially connected HBF has the potential to provide an good compromise between hardware complexity and system performance in comparison to fully digital beamforming.

Breast cancer is the most prevalent neoplasia among women, with early and accurate diagnosis critical for effective treatment. In clinical practice, however, the subjective nature of histological grading of infiltrating ductal adenocarcinoma of the breast (DAC-NOS) often leads to inconsistencies among pathologists, posing a significant challenge to achieving optimal patient outcomes. Our study aimed to address this reproducibility problem by leveraging artificial intelligence (AI). We trained a deep-learning model using a convolutional neural network-based algorithm (CNN-bA) on 100 whole slide images (WSIs) of DAC-NOS from the Cancer Genome Atlas Breast Invasive Carcinoma (TCGA-BRCA) dataset. Our model demonstrated high precision, sensitivity, and F1 score across different grading components in about 17.5 h with 19,000 iterations. However, the agreement between the model's grading and that of general pathologists varied, showing the highest agreement for the mitotic count score. These findings suggest that AI has the potential to enhance the accuracy and reproducibility of breast cancer grading, warranting further refinement and validation of this approach.

The activity-regulated cytoskeleton-associated (Arc) protein is essential for synaptic plasticity and memory formation. The Arc gene, which contains remnants of a structural GAG retrotransposon sequence, produces a protein that self-assembles into capsid-like structures harboring Arc mRNA. Arc capsids, released from neurons, have been proposed as a novel intercellular mechanism for mRNA transmission. Nevertheless, evidence for intercellular transport of Arc in the mammalian brain is still lacking. To enable the tracking of Arc molecules from individual neurons in vivo, we devised an adeno-associated virus (AAV) mediated approach to tag the N-terminal of the mouse Arc protein with a fluorescent reporter using CRISPR/Cas9 homologous independent targeted integration (HITI). We show that a sequence coding for mCherry can successfully be knocked in at the 5' end of the Arc open reading frame. While nine spCas9 gene editing sites surround the Arc start codon, the accuracy of the editing was highly sequence-dependent, with only a single target resulting in an in-frame reporter integration. When inducing long-term potentiation (LTP) in the hippocampus, we observed an increase of Arc protein highly correlated with an increase in fluorescent intensity and the number of mCherry-positive cells. By proximity ligation assay (PLA), we demonstrated that the mCherry-Arc fusion protein retains the Arc function by interacting with the transmembrane protein stargazin in postsynaptic spines. Finally, we recorded mCherry-Arc interaction with presynaptic protein Bassoon in mCherry-negative surrounding neurons at close proximity to mCherry-positive spines of edited neurons. This is the first study to provide support for inter-neuronal in vivo transfer of Arc in the mammalian brain.

The decrease in catch of valuable fish species from natural water bodies is compensated by their intensive cultivation in artificial conditions. In order to realize the main task of trout farms associated with obtaining marketable products in the shortest possible period of time, artificial feed is used as a source of food. A economically feasible alternative source of raw materials is the products of plant origin. In the process of production of oil and fat products at various stages, numerous fat wastes and by-products are formed, wich have fodder value and are not used as feeding facilities in industrial scale. This is especially frue for fat processing (soapstock of Light oil, fatty bleaching clays, dezodoration chases, phosphatides, calcium salts of fatty acids), as well as waste oils in combination with fat processing waste. On the base of studying the classical technology of producing mixed fodders for valuable fish species and eliminating its drawbacks the technology of pseudocapsulated mixed fodders for salmon fish grown in the CFD of Russia Federation with given fodder value and a line for its realization is suggested. The best way to bring fats and vitamins contained in them to valuable fish species is to feed them as a part of mixed fodders. At present, however, it is difficult to increase the level of fat in feed on the existing Russian fats supply lines, since most plants can include up to 10% of fat in the bulk feed line. Therefore, the development of mixing formulations and techniques for valuable fish breeds using sturgeon fish with more than 10% fat in it as an addition of fat oily wastes such as epaulettes and phosphates is not only of scientific interest. In the course of studies optimization encapsulated optimal feeds for program "Feed Optima Expert" pseudorainbow trout were also developed. As a result, close results were obtained in terms of the rate of growth of trout, the development, conversion of feed and viability in comparison with the best feed counterparts on the Russian market.

The incidence of non-alcoholic fatty liver disease (NAFLD) is a continuously growing health problem worldwide, along with obesity. Therefore, both novel methods to efficiently study the manifestation of NAFLD and to analyze drug efficacy in pre-clinical models are needed. In the present study, we developed a deep neural network -based model to quantify micro- and macrovesicular steatosis in the liver on hematoxylin-eosin stained whole slide images (WSIs), using the cloud-based platform, Aiforia Create (Aiforia Technologies, Helsinki, Finland). The training data included a total of 101 WSIs from dietary interventions of wild-type mice and from two genetically modified (GM) mouse models with steatosis. The algorithm was trained for the following: to detect liver parenchyma, to exclude the blood vessels and any artefacts generated during tissue processing and image acquisition, to recognize and differentiate the areas of micro- and macrovesicular steatosis, and to quantify the recognized tissue area. The results of the image analysis replicated well the evaluation by expert pathologists, and correlated well with the liver fat content measured by EcoMRI ex vivo, and the correlation with total liver triglycerides were notable. In conclusion, the developed deep learning-based model is a novel tool for studying liver steatosis in mouse models on paraffin sections, and thus, can facilitate reliable quantification of the amount of steatosis in large preclinical study cohorts.