Accelerated Nuclear Magnetic Resonance Spectroscopy with Deep Learning

Nuclear magnetic resonance (NMR) spectroscopy serves as an indispensable tool in chemistry and biology but often suffers from long experimental times. We present a proof‐of‐concept of the application of deep learning and neural networks for high‐quality, reliable, and very fast NMR spectra reconstruction from limited experimental data. We show that the neural network training can be achieved using solely synthetic NMR signals, which lifts the prohibiting demand for a large volume of realistic training data usually required for a deep learning approach.     

A modified quick,easy, cheap,effective, rugged,and safe method with hydrophobic natural deep eutectic solvent as extractant and analyte protectant for analyzing pyrethroid residues in tomatoes

In this work, a novel quick, easy, cheap, effective, rugged, and safe technique with hydrophobic natural deep eutectic solvent as both extractant and analyte protectant was developed and combined with gas chromatography–tandem mass spectrometry to analyze pyrethroid residues in tomatoes. Eight hydrophobic natural deep eutectic solvents were first evaluated as analyte protectants and those with decanoic acid or lactic acid as hydrogen bond donor were demonstrated to be effective in compensating for the matrix effects of pyrethroids in the gas chromatography system. Hence, they were added to solvent standards for correcting the quantitation errors instead of matrix‐matched calibration standards. Then the abilities of these acid‐based deep eutectic solvents to extract pyrethriods from tomatoes were evaluated. Results showed the recoveries of all pyrethroids reached to over 80% with only 5 mL menthol:decanoic acid (1:1) used, and good phase separation was easily achieved without the addition of inorganic salt in the extraction step, indicating hydrophobic natural deep eutectic solvent could be a green substitute for acetonitrile in the quick, easy, cheap, effective, rugged, and safe extraction. Compared with the conventional method, the proposed protocol improved the recoveries, reduced the matrix effects, and simplified the extraction step, demonstrating to be an effective, fast, and green method.     

A three‐phase solvent extraction system combined with deep eutectic solvent‐based dispersive liquid–liquid microextraction for extraction of some organochlorine pesticides in cocoa samples prior to gas chromatography with electron capture detection

A sample pretreatment method based on the combination of a three‐phase solvent extraction system and deep eutectic solvent‐based dispersive liquid–liquid microextraction has been introduced for the extraction of four organochlorine pesticides in cocoa samples before their determination by gas chromatography‐electron capture detection. A mixture of sodium chloride, acetonitrile, and potassium hydroxide solution is added to cocoa bean or powder. After vortexing and centrifugation of the mixture, the collected upper phase (acetonitrile) is removed and mixed with a few microliters of N,N‐diethanol ammonium chloride: pivalic acid deep eutectic solvent. Then it is rapidly injected into deionized water and a cloudy solution is obtained. Under optimum conditions, the limits of detection and quantification were found to be 0.011‐0.031 and 0.036‐0.104 ng/g, respectively. The obtained extraction recoveries varied between 74 and 92%. Also, intra‐ (n = 6) and interday (n = 4) precisions were less than or equal to 7.1% for the studied pesticides at a concentration of 0.3 ng/g of each analyte. The suggested method was applied to determine the studied organochlorine pesticide residues in various cocoa powders and beans gathered from groceries in Tabriz city (Iran) and aldrin and dichlobenil were found in some of them.     

Combining convolutional neural networks and in‐line near‐infrared spectroscopy for real‐time monitoring of the chromatographic elution process in commercial production of notoginseng total saponins

The chromatographic elution process is a key step in the production of notoginseng total saponins. Due to quality variability of loading samples and resin capacity decreasing over cycle time, saponins, especially the five main saponins of notoginseng total saponins, need to be monitored in real time during the elution process. In this study, convolutional neural networks, one of the most popular deep learning methods, were used to develop quantitative calibration models based on in‐line near‐infrared spectroscopy for notoginsenoside R₁, ginsenosides Rg₁, Re, Rb₁ and Rd, and their sum concentration, with root mean square error of prediction values of 0.87, 2.76, 0.60, 1.57, 0.28, and 4.99 mg/mL, respectively. Partial least squares calibration models were also developed for model performance comparison. Results show predicted concentration profiles outputted by both the convolutional neural network models and partial least squares models show agreements with the real trends defined by reference measurements, and can be used for elution process monitoring and endpoint determination. To the best of our knowledge, this is the first reported case study of combining convolutional neural networks and in‐line near‐infrared spectroscopy for monitoring of the chromatographic elution process in commercial production of botanical drug products.     

Development of a stir bar sorptive extraction method coupled to solidification of floating droplets dispersive liquid–liquid microextraction based on deep eutectic solvents for the extraction of acidic pesticides from tomato samples

A stir bar sorptive extraction method coupled with deep eutectic solvent based solidification of floating organic droplets–dispersive liquid–liquid microextraction has been used for the simultaneous derivatization and extraction of some acidic pesticides in tomato samples. In this method, initially the analytes are adsorbed on a coated stir bar from tomato juice filled in a narrow tube. After extraction, the stir bar is removed and a water–miscible deep eutectic solvent is used to elute the analytes. Afterward, a derivatization agent and a water–immiscible deep eutectic solvent (as an extraction solvent) with melting point near to room temperature are added to the obtained eluant at µL–levels and the obtained mixture is rapidly injected into deionized water. Under the optimum conditions, the introduced method indicated high enhancement (1543–3353) and enrichment (2530–2999) factors, low limits of detection (7–14 ng/L) and quantification (23–47 ng/L), good linearity (r² ≥ 0.9982), and satisfactory repeatabilities (relative standard deviation ≤12% for intra– and inter–day precisions at a concentration of 100 ng/L of each analyte). Finally, the proposed method was applied in analysis of the analytes in tomato samples.     

Ligandless,deep eutectic solvent‐based ultrasound‐assisted dispersive liquid‐liquid microextraction with solidification of the aqueous phase for preconcentration of lead,cadmium, cobalt and nickel in water samples

A green and efficient sample preparation method using a deep eutectic solvent‐based ultrasounds‐assisted dispersive liquid–liquid microextraction with solidification of the aqueous phase followed by high performance liquid chromatography analysis was developed for preconcentration and determination of heavy metals in environmental samples. In the proposed method, a novel, low density deep eutectic solvent was prepared by mixing trihexyl(tetradecyl)phosphonium chloride and thiosalicylic acid at a molar ratio of 1:2 and used both as an extractant and complexing agent. Ultrasound was used to disperse the extractant in the aqueous phase of the sample. Then, the phases were separated by centrifugation, after which the aqueous phase was frozen and the surface extractant phase was dissolved in a small volume of acetonitrile and subjected to liquid chromatographic analysis. The proposed method provided precisions (relative standard deviation, n = 5) in the range of 2.6–4.7%. The limit of detection were 0.05, 0.13, 0.06, and 0.11 µg/L for Pb(II), Cd(II), Co(II), Ni(II), respectively. The enhancement factors were equal to 154, 159, 162, and 158 for lead(II), cadmium(II), cobalt(II), and nickel(II), respectively. The accuracy of the proposed method was evaluated using certified reference materials (CA011b – hard drinking water, NIST 1643e – trace elements in water, TMRAIN‐04 – simulated rain sample).     

医学影像处理的深度学习可解释性研究进展

随着医学影像数据的迅速增长,传统的影像分析方法给医生带来巨大挑战。利用计算机视觉技术提供自动或半自动辅助诊断,可大大缓解人工阅片压力,提高诊断的准确性,促进医疗流程的标准化建设等。目前,深度学习卷积神经网络在医学影像处理中已取得不俗表现,但深度学习“黑匣子”的不可解释性阻碍了智能医疗诊断的发展。为增强对医学影像数据处理的深度学习可解释性的了解,对近几年相关研究进展进行了综述。首先,综述了深度学习在医学领域的应用现状及面临的问题,对神经网络的可解释性内涵进行了讨论;然后,从现有深度学习可解释性的常见方法出发,重点讨论了医学影像处理的深度学习可解释性研究进展;最后,探讨了医学影像处理的深度学习可解释性的发展趋势。     

游客情感计算的文本大数据挖掘方法比较研究

旅游文本大数据以其方便、快捷和低门槛的特点为游客情感计算提供了极大便利,已经成为旅游大数据的主要来源之一。基于大数据理论和情感理论,以文本大数据为数据源,在全面梳理国内外情感计算相关成果的基础上,利用人工智能中的逻辑/算法编程方法、机器学习方法、深度学习方法对旅游文本大数据进行挖掘,探索最佳的基于文本大数据的游客情感计算方法。研究发现：（1）基于情感词典的游客情感计算模型,其核心是构建情感词典和设计情感计算规则,方法简单,容易实现,适用语料范围广。（2）机器学习,用统计学方法抽取文本中的特征项,具有非线性特征,可靠性较线性特征的情感词典方法高。（3）基于深度学习技术的游客情感计算,效果良好,准确率在85%以上。训练多领域的文本语料易于移植,实用性强,且泛化能力好,较适合大数据时代游客情感计算研究。     

融合图卷积神经网络和注意力机制的PM2.5小时浓度多步预测

PM_2.5小时浓度多为单步预测。为实现PM_2.5小时浓度的多步预测,基于“编码器-解码器”的序列-序列预测（Seq2Seq）模型,集合图卷积神经网络提取非欧式空间数据特征的能力以及注意力机制自适应关注特征的能力,提出了融合图卷积神经网络和注意力机制的PM_2.5小时浓度多步预测（GCN_Attention_Seq2Seq）模型。并与Seq2Seq模型和使用了图卷积神经网络、未使用注意力机制的GCN_Seq2Seq模型进行了对照,以2015—2016年北京市22个空气质量监测站点的空气质量数据为样本进行实例验证,结果表明,Seq2Seq模型和图卷积神经网络（GCN）可对PM_2.5小时浓度数据的时空依赖进行有效建模,注意力机制有助于减缓多步预测中的预测精度衰减,提升PM_2.5小时浓度多步预测的精度。GCN_Attention_Seq2Seq模型可有效应用于多种长度的PM_2.5浓度预测窗口。     

Application of Deep Eutectic Solvents in the Synthesis of Substituted 2-Mercaptoquinazolin-4(3H)-Ones: A Comparison of Selected Green Chemistry Methods

In this study, deep eutectic solvents (DESs) were used as green and eco-friendly media for the synthesis of substituted 2-mercaptoquinazolin-4(3H)-ones from different anthranilic acids and aliphatic or aromatic isothiocyanates. A model reaction on anthranilic acid and phenyl isothiocyanate was performed in 20 choline chloride-based DESs at 80 °C to find the best solvent. Based on the product yield, choline chloride:urea (1:2) DES was found to be the most effective, while DESs acted both as solvents and catalysts. Desired compounds were prepared with moderate to good yields using stirring, microwave-assisted, and ultrasound-assisted synthesis. Significantly, higher yields were obtained with mixing and ultrasonication (16–76%), while microwave-induced synthesis showed lower effectiveness (13–49%). The specific contribution of this research is the use of DESs in combination with the above-mentioned green techniques for the synthesis of a wide range of derivatives. The structures of the synthesized compounds were confirmed by ¹H and ¹³C NMR spectroscopy.