Machine learning algorithm to predict mortality in critically ill patients with sepsis-associated acute kidney injury.

This study aimed to establish and validate a machine learning (ML) model for predicting in-hospital mortality in patients with sepsis-associated acute kidney injury (SA-AKI). This study collected data on SA-AKI patients from 2008 to 2019 using the Medical Information Mart for Intensive Care IV. After employing Lasso regression for feature selection, six ML approaches were used to build the model. The optimal model was chosen based on precision and area under curve (AUC). In addition, the best model was interpreted using SHapley Additive exPlanations (SHAP) values and Local Interpretable Model-Agnostic Explanations (LIME) algorithms. There were 8129 sepsis patients eligible for participation; the median age was 68.7 (interquartile range: 57.2-79.6) years, and 57.9% (4708/8129) were male. After selection, 24 of the 44 clinical characteristics gathered after intensive care unit admission remained linked with prognosis and were utilized developing ML models. Among the six models developed, the eXtreme Gradient Boosting (XGBoost) model had the highest AUC, at 0.794. According to the SHAP values, the sequential organ failure assessment score, respiration, simplified acute physiology score II, and age were the four most influential variables in the XGBoost model. Individualized forecasts were clarified using the LIME algorithm. We built and verified ML models that excel in early mortality risk prediction in SA-AKI and the XGBoost model performed best.

A convolutional recurrent neural network with attention framework for speech separation in monaural recordings.

Most speech separation studies in monaural channel use only a single type of network, and the separation effect is typically not satisfactory, posing difficulties for high quality speech separation. In this study, we propose a convolutional recurrent neural network with an attention (CRNN-A) framework for speech separation, fusing advantages of two networks together. The proposed separation framework uses a convolutional neural network (CNN) as the front-end of a recurrent neural network (RNN), alleviating the problem that a sole RNN cannot effectively learn the necessary features. This framework makes use of the translation invariance provided by CNN to extract information without modifying the original signals. Within the supplemented CNN, two different convolution kernels are designed to capture information in both the time and frequency domains of the input spectrogram. After concatenating the time-domain and the frequency-domain feature maps, the feature information of speech is exploited through consecutive convolutional layers. Finally, the feature map learned from the front-end CNN is combined with the original spectrogram and is sent to the back-end RNN. Further, the attention mechanism is further incorporated, focusing on the relationship among different feature maps. The effectiveness of the proposed method is evaluated on the standard dataset MIR-1K and the results prove that the proposed method outperforms the baseline RNN and other popular speech separation methods, in terms of GNSDR (gloabl normalised source-to-distortion ratio), GSIR (global source-to-interferences ratio), and GSAR (gloabl source-to-artifacts ratio). In summary, the proposed CRNN-A framework can effectively combine the advantages of CNN and RNN, and further optimise the separation performance via the attention mechanism. The proposed framework can shed a new light on speech separation, speech enhancement, and other related fields.

Separation and purification of lanosterol, dihydrolanosterol, and cholesterol from lanolin by high-performance counter-current chromatography dual-mode elution method.

Lanosterol is a potential drug for cataracts treatment, which can reverse the aggregation of intracrystalline proteins. The low concentration in lanolin calls for high-performance separation methods. In this study, a counter-current chromatography dual-mode elution method was developed for the first time to separate and purify lanosterol from hexane extract of lanolin after saponification, in which the column was first eluted with the lower phase as mobile phase in head-to-tail mode, followed by the upper phase in the tail-to-head mode. High purity of lanosterol, dihydrolanosterol, and cholesterol can be obtained simultaneously. A solvent system composed of n-heptane/acetonitrile/ethyl acetate (5:5:1, v/v/v) was selected and optimized via partition coefficient determination. Compounds such as 111 mg lanosterol, 84 mg dihydrolanosterol, and 183 mg cholesterol with high purity of 99.77, 95.71, and 91.43%, respectively, analyzed by high-performance liquid chromatography were obtained within 80 min from 700 mg crude extract from 1.78 g lanolin. The method was also used to improve the purity of commercial lanosterol product from 66.97 to above 99%. Counter-current chromatography could serve as a potential and powerful technique for commercial production of highly pure lanosterol.

Identification of two novel TPK1 gene mutations in a Chinese patient with thiamine pyrophosphokinase deficiency undergoing whole exome sequencing.

Background The mutations of thiamine pyrophosphokinase-1 (TPK1) gene have been frequently studied in some patients with thiamine metabolism dysfunction syndrome-5 (THMD5), while TPK1 mutations in Chinese patients have been investigated by only homozygous. A search of the literature on the mutations in the Chinese population currently published revealed that no reports of compound heterozygous mutations were reported. Here, we report a Chinese patient with compound heterozygous TPK1 mutations who underwent magnetic resonance imaging (MRI), whole exome sequencing (WES), molecular diagnosis, bioinformatics analysis, and three-dimensional (3D) protein structure analysis. Case presentation A Chinese boy was born after an uneventful pregnancy to non-consanguineous and healthy parents. On the sixth day after his birth, the lactate level of the patient was between 8.6 mmol/L and 14.59 mmol/L in plasma (the normal level is in the range of 0.5-2.2 mmol/L). Lactate was reduced to the normal level after rehydration, acid correction, expansion, and other treatments. After 4 months, the patient presented with an acute, 3-h-long, non-induced convulsions, and was admitted to our hospital for weakness, decreased oral intake, and lethargy. Results achieved by electroencephalography (EEG), cerebrospinal fluid, and other biochemical findings were normal. A visible hemorrhagic lesion was also observed in the brain. Seizures increased significantly during infection, which was accompanied by higher lactic acid levels. MRI of the brain showed an obvious signal shadow, in which bilateral frontal and temporal parietal subarachnoid cavities were widened, and more abnormal signals were observed; therefore, further consideration of hypoxic-ischemic encephalopathy and genetic metabolic disease was taken into account. Conclusions The results of WES revealed that the patient was associated with compound heterozygous mutations NM_022445.3:c.[263G>A]; [226A>G] of TPK1. His parents were non-consanguineous; while his father was found to be a heterozygous carrier with the mutation c.[263G>A], his mother was identified as a heterozygous carrier with the mutation c.[226A>G]. The results indicated that the patient had a compound heterozygous TPK1 mutation, and this is the first reported case in China.

Insights into the cellular function of YhdE, a nucleotide pyrophosphatase from Escherichia coli.

YhdE, a Maf-like protein in Escherichia coli, exhibits nucleotide pyrophosphatase (PPase) activity, yet its cellular function remains unknown. Here, we characterized the PPase activity of YhdE on dTTP, UTP and TTP and determined two crystal structures of YhdE, revealing 'closed' and 'open' conformations of an adaptive active site. Our functional studies demonstrated that YhdE retards cell growth by prolonging the lag and log phases, particularly under stress conditions. Morphology studies showed that yhdE-knockout cells transformed the normal rod shape of wild-type cells to a more spherical form, and the cell wall appeared to become more flexible. In contrast, YhdE overexpression resulted in filamentous cells. This study reveals the previously unknown involvement of YhdE in cell growth inhibition under stress conditions, cell-division arrest and cell-shape maintenance, highlighting YhdE's important role in E. coli cell-cycle checkpoints.

[Effects and the mechanisms of cardiac short-term memory on cellular electrical excitability].

Electrical instability easily induces a unidirectional conduction block, resulting in ventricular tachycardia (VT) or even fibrillation (VF). Cardiac memory affects dynamic electrical characteristics through previous pacing so that it makes the memory important in arrhythmia study. This paper investigates the impact of the rapid pacing duration on cellular excitability and its mechanism. Based on the canine endocardial single cell, a one-dimensional tissue model was developed. Simulations were realized with OpenMP parallel programming method. The results showed that with repetitive pacing, the cellular excitability became low while the conduction velocity decreased. Accumulation of intracellular [Ca2+]i and [Na+]i and depletion of [K+]i led to the shift of membrane current-voltage curves, changing the membrane resistance. Excitability determined by the resistance at the large width of stimulus pulse, therefore, it suggested that [Ca2+]i and [K+]i-induced memory formed the ionic substrates for the alteration of excitability.

Short-term memory and electrical restitution in the canine transmural ventricle.

Cardiac short-term memory is an intrinsic property of paced myocardium that reflects the influence of pacing history. Using an optical mapping method to record membrane voltage and intracellular calcium (Ca(2 +)(i)), this study investigated the properties and mechanisms of short-term memory in isolated and perfused canine wedge preparations. In addition to the dynamic and S1S2 pacing protocols, a perturbed downsweep pacing protocol was used to get a complete overview of the restitution portrait. Abrupt changes in basic cycle length (BCL) were applied to investigate the accommodation process of action potential duration (APD). The results showed unobvious differences of memory among the epi-, mid- and endo-myocytes, implying an insignificant memory-induced transient heterogeneity in APD across the transmural canine hearts. With the decrease of pacing rate S1, memory gradually elevated and achieved a maximum around 400 ms, and then reduced as S1 decreased further, indicating a non-monotonic relationship between memory and the pacing rate. After suppressing the Ca(2 +)(i) transient with ryanodine (3 µmol l(-1)), the accommodation process of APD to a new BCL significantly abbreviated (τ = 37.41 ± 4.42 stimuli before ryanodine, τ = 15.84 ± 4.74 stimuli after ryanodine, p < 0.01). Therefore, Ca(2 +)(i) cycling was suggested to play an important role in memory during dynamic pacing.

Vulnerability in simulated ischemic ventricular transmural tissues.

Vulnerability is an effective index to evaluate increased risk for unidirectional conduction block and reentry in hearts. Recent reports in animal experiments have indicated an opposite characteristics of the vulnerability in normal and ischemic transmural tissues. In order to clarify the differences and to investigate the mechanisms, a computer simulation method was used in this study to investigate the vulnerability relative to the premature pacing sites in normal and ischemic transmural tissues. Endo-, mid- and epi-cardial myocytes incorporating different severities of ischemia were developed across a tissue strand. The sodium channel inactivation gating variable h was calculated to provide the degree of sodium current recovery preceding the premature pacing. In the normal tissue, the measured vulnerable window was demonstrated to be wider by delivering an endocardial premature beat than that by applying an epicardial premature pacing. On the contrary, during ischemia the epicardium showed a wider vulnerable window than the endocardium. The results illustrated that during ischemia h decreased with accumulation of [K⁺]o, and action potential duration dispersion was obviously altered due to anoxia. In contrast, the elevated [K⁺]o was suggested to play an important role in the difference of the location-dependent vulnerability in normal and ischemic tissues.