-
抗生素是由微生物(包括细菌、真菌、放线菌属)或高等动植物所产生的,能干扰其他生物细胞发育功能的化学物质[1]。常见的抗生素有青霉素类、四环素类、头孢类、喹诺酮类等,其抑菌性被广泛应用于医疗卫生、畜牧业和水产养殖业[2]。然而进入生物体内的抗生素不能被完全代谢,残留的抗生素可能会通过市政污水、牲畜粪便、农田径流等多种途径排入环境中[3-5]。近年来,已经在地表水、污水处理厂、土壤和大气中检测到抗生素的存在[6-9]。环境中抗生素的存在会对人体健康和环境产生不利影响,如增加环境中微生物群落的选择压力,临床抗生素使用效率降低等[5, 10]。系统分析环境中残留抗生素的种类、明确其与微生物的作用机制以及开展不同处理方法的效果分析,对于准确评估抗生素的风险和控制其不利影响具有重要意义。
目前抗生素的定性和定量分析方法较多,包括分光光度法、化学荧光、化学磷光、液相色谱法和毛细管电泳法等,这些方法耗时长、检测过程复杂[11],对大量检测数据进行有效的分析显得尤为重要。环境中残留的抗生素与微生物的作用机制复杂,高效的数据处理方法如机器学习算法有助于深入揭示作用机制。鉴于环境中残留抗生素的潜在风险,研究者发现膜过滤、高级氧化技术、生物处理、吸附[12, 13]、光催化等技术均可实现抗生素的去除[14-16]。然而多种处理方法对于不同环境介质中不同种类抗生素去除效果的评估、优化以及预测都有待进一步探究,建立相关的模型来分析去除方法、目标污染物和去除率之间的关系对于抗生素的污染控制具有指导意义。
机器学习是一种旨在使计算机具有获取知识和处理大数据的能力,同时建立学习理论计算方法,构建各种学习系统并将其投入到实际应用中的一种技术[17]。其核心是“使用算法解析数据,从中学习,然后对新数据做出决定或预测”。机器学习算法可以分为有监督、无监督和半监督[18],主要包括决策树学习、朴素贝叶斯、支持向量机、随机森林、k-均值算法、主成分分析、人工神经网络、k-近邻分析和遗传算法等[19]。鉴于机器学习可以处理大量样本数据,并且能进行图像识别和语言识别,目前在大气污染物的检测和评估以及医疗中抗生素的选择和适用中已得到较为成熟的发展[20-22]。在抗生素的相关研究中,传统分析方法与机器学习的结合可以更加快速的识别抗生素的种类并进行定量分析[23],并且利用机器学习优化模型有助于最佳去除条件的获取。
基于此,本文主要综述了:(1)机器学习作为一种辅助手段在抗生素鉴定识别中的应用,主要包括食品和环境中抗生素的定性和定量分析以及新型抗生素的发现;(2)机器学习在抗生素与微生物作用机制研究中的应用;(3)机器学习模型用于抗生素去除效果评估。本文主要通过综述机器学习算法在抗生素鉴定识别、微生物作用机制和去除效果评估预测方面的应用现状,针对不同算法的特点和局限性,为今后的研究方向和发展提出展望。
机器学习在抗生素的鉴定识别、微生物作用机制以及去除效果评估中的应用研究进展
Research progress in the application of machine learning in the identification of antibiotics, microbial mechanism of action and evaluation of removal effect
-
摘要: 抗生素在医疗卫生、养殖业等领域的广泛应用导致其随着医疗废弃物、废水等进入到自然环境并对人体健康和生态环境造成不利影响,而系统分析环境中残留抗生素的种类、明确其与微生物的作用机制以及开展不同处理方法的效果分析,对于准确评估抗生素的风险和控制其不利影响具有重要意义。作为一种辅助手段,机器学习算法在大量数据解析的基础上可进行结果评估和预测,因此可高效、低成本分析环境污染物的行为特征。基于此,本文综述了机器学习算法在抗生素鉴定识别、微生物作用机制和去除效果评估预测方面的应用现状,并概括了不同算法的应用特点和局限性。鉴于机器学习当前在抗生素研究中的重要作用,为其未来研究方向和发展提出展望,如在其它新兴污染物的环境行为、效应及控制等方面的应用。Abstract: The wide application of antibiotics in medical and health,breeding and other fields leads to their entry into the natural environment along with medical waste and waste water. It causes adverse effects on human health and ecological environment. It is of great significance for accurately assessing the risk of antibiotics and controlling their adverse effects to analyze the types of residual antibiotics in the environment, the mechanism of their interaction with microorganisms and the effect analysis of different treatment methods. As an auxiliary method, machine learning algorithms can evaluate and predict the results based on the analysis of large amounts of data, which can effectively and cheaply analyze the behavioral characteristics of environmental pollutants. Based on this, this paper reviewed the application status of machine learning algorithms in antibiotic identification, microbial mechanism of action and evaluation of removal effect. The application characteristics and limitations of different algorithms are also summarized. In view of the important role of machine learning in the current antibiotic research, the future research direction and development of machine learning are proposed, such as its application in the environmental behavior, effect and control of other emerging pollutants.
-
表 1 不同机器学习方式应用于抗生素定量分析的性能比较
Table 1. Comparison of performance of different machine learning methods applied to quantitative analysis of antibiotics
机器学习算法
Machine learning algorithm抗生素测定方法
Methods for determination of antibiotics样品来源
Source of samples抗生素种类
Kinds of Antibiotics分析效果
Effects of analysis参考文献
References人工神经网络 基于数字图像的比色法 牛奶 四环素类 可测浓度范围为1.0—9.0 µg·mL−1,平均回收百分比
>90%,均方差<0.2%,相对准确度:91.9%—105%. 可缩短分析时间,减少溶剂使用[28, 33] 采用统计计量方法对短波红外光谱进行预处理 药物 乙酰螺旋
霉素粉末标准正态变量预处理后建立的人工神经网络的R2最大值>0.997
均方根误差最小值:3.677×10−3多元线性回归 太赫兹光谱分析 饲料基质、
食品氟喹诺酮类 诺氟沙星、恩诺沙星和氧氟沙星的最佳相关系数:0.867、0.828和0.964
太赫兹光谱分析可用于单个抗生素的定量分析,但是对于混合抗生素的检测有一定的局限性[27] 线性判别分析和多层感知器神经网络 电子鼻装置 山羊乳样品 青霉素G 线性判别分析准确率:训练集98%,测试集96.7%
多层感知器神经网络准确率:训练集97%,测试
集:94.9%[32] 主成分人工神经
网络方波伏安法 自制生物
溶液四环素和
头孢克肟四环素检测范围:10−5—10−3 mol·L−1
头孢克肟检测范围:10−5—10−3 mol·L−1
拟合度较好
3种介质中抗生素检测R2>0.972,平均回收率92.3%—113%,峰值电流可重复性<10%[29] 偏最小二乘法和多层前馈神经网络 荧光光谱 经不同预处理的牛奶 磺胺噻唑 多层前反馈神经网络的校准的准确性比偏最小二乘法高,在最佳设计条件下,对于磺胺噻唑的回收率可以达到88% [34] 主成分回归和径向基函数人工神经
网络差分脉冲溶出伏
安法牛奶、蜂蜜、鸡蛋 左氧氟沙星、加替沙星和洛美沙星 主成分分析预测能力优于径向基函数神经网络,适用于复杂的线性分析系统
主成分分析相对预测误差<10%
鸡蛋和蜂蜜中洛美沙星的回收率<80%
其余食品样品中回收率:99%—110.7%[24] 表 2 不同机器学习算法识别抗菌肽的性能
Table 2. The properties of antimicrobial peptides identified by different machine learning algorithm
机器学习算法
Machine learning algorithm灵敏度/%
Sensitivity特异性/%
Specificity准确度/%
Accuracy马修斯相关系数a
Matthews correlation coefficient受试者操作特征曲线下面积/%
AUROC参考文献
References人工神经网络 82.98 85.09 84.04 0.6809 84.06 [37, 42] 判别分析 87.08 80.76 83.92 0.6797 89.97 [37] 随机森林 92.70 82.44 87.57 0.7554 93.63 [37] 支持向量机 88.90 79.92 84.41 0.6910 90.63 [37] 基于长期短期记忆网络的
深度神经网络89.89 92.13 91.01 0.8204 96.48 [44] 基于SMO算法的支持向量机 88.5 80 94 0.76 — [40] 基于N端和C端残基的定量矩阵 90.02 90.72 90.37 — — [43] 基于N端和C端残基的支持向量机 92.11 92.11 92.11 — — [43] 基于N端和C端残基的人
工神经网络88.17 88.17 88.17 — — [43] 极端梯度增强树 >97.27 >92.29 — >0.89 >98 [46] 卷积神经网络 >76.92 >68.18 >94.18 >0.70 — [46] a马修斯相关系数;用以测量二分类的分类性能的指标,它的取值范围为[−1,1],取值为1时表示对受试对象的完美预测,取值为0时表示预测的结果不如随机预测的结果,-1是指预测分类和实际分类完全不一致。
Matthews correlation coefficient:An index used to measure the classification performance of a dichotomy. It's in the range of [-1,1], a value of 1 indicates a perfect prediction of the subject, 0 indicates that the predicted result is worse than the random predicted result, -1 is when the predicted classification is completely inconsistent with the actual classification.表 3 机器学习在抗生素去除中的应用及模型相关系数
Table 3. Application of machine learning in antibiotic removal in the environment and model correlation coefficient
机器学习算法
Machine learning algorithm去除方法
Removal method抗生素种类
Kinds of Antibiotics相关系数R2
Correlation index参考文献
References人工神经网络 膨润土和活性炭吸附 环丙沙星 概率预处理后
膨胀土R2=0.998
活性炭R2=0.9919[68] 鸡粪和甘蔗堆肥 四环素 >0.973 [65] 纳米粒子光降解 头孢克肟 0.996 [15] TiO2悬浮液光催化 土霉素 0.99751 [14] 生物吸附 四环素 最优解可达到0.99984 [64] 太阳能-芬顿工艺 四环素 不同反应条件下均大于0.97,最高可达1 [69] 臭氧高级氧化 磺胺类抗生素 0.97 [70] 磁化金属有机骨架 沙拉沙星 0.9951 [71] Ag/TiO2纳米片/还原石墨烯光催化 四环素 最高可达0.99894 [62, 66] 固定床和浆液光催化 奥硝唑 检验R2值=0.946 [67] 超滤 恩诺沙星 基于膜通透性R=0.99036
基于截留率R=0.9756[72] 污泥和活性炭吸附 恩诺沙星 0.999 [73] 亚临界水氧化法 替卡西林 0.9861 [74] 零价双金属纳米还原 氯霉素 0.999 [75] 反向传播神经
网络厌氧膜生物反应器 β-内酰胺类抗生素 >0.968 [76] 分子组装多孔氮化碳光催化 磺胺嘧啶 0.9556 [77] 径向基函数神经网络 内置上流式厌氧污泥包层生物电化学系统,
不同水力停留时间下处理β-内酰胺类抗生素 水力停留时间为72 h时取得最优解,R2=0.979 [78] 物理降解 苏云金素 >0.996 [59] 遗传算法-人工神经网络 光催化、光芬顿 甲硝唑 >0.9877 [63] 随机森林 碳基材料吸附 四环素和磺胺甲恶唑 四环素:0.894
磺胺甲恶唑:0.909[60] 响应曲面法 污泥和活性炭吸附 恩诺沙星 0.995 [73] 磁化金属有机骨架 沙拉沙星 0.9995 [71] 纳米复合材料吸附 四环素 0.9668 [61] 臭氧高级氧化 磺胺类抗生素 0.99 [70] 亚临界水氧化 替卡西林 0.9775 [74] 零价双金属纳米粒子还原 氯霉素 0.996 [75] 增强回归树 纳米复合材料吸附 四环素 0.9992 [61] 一般回归神经网络 纳米复合材料吸附 四环素 0.9920 [61] -
[1] 张晓蕾. 对antibiotics由抗菌素到抗生素之分析 [J]. 临床医药文献电子杂志, 2014, 1(11): 2061-2062. ZHANG X L. Analysis of antibiotics from antimicrobial to antibiotics [J]. Journal of Clinical Medical Literature, 2014, 1(11): 2061-2062(in Chinese).
[2] 廖洋, 鲁金凤, 曹轶群, 等. 光催化降解对抗生素藻类毒性效应影响研究进展 [J]. 环境化学, 2021, 40(1): 111-120. doi: 10.7524/j.issn.0254-6108.2019122404 LIAO Y, LU J F, CAO Y Q, et al. Research progress on the effects of photocatalytic degradation on the algae toxicity of antibiotics [J]. Environmental Chemistry, 2021, 40(1): 111-120(in Chinese). doi: 10.7524/j.issn.0254-6108.2019122404
[3] LACH J, STĘPNIAK L, OCIEPA-KUBICKA A. Antibiotics in the environment as one of the barriers to sustainable development [J]. Problemy Ekorozwoju, 2018, 13(1): 197-207. [4] REINTHALER F F, POSCH J, FEIERL G, et al. Antibiotic resistance of E. coli in sewage and sludge [J]. Water Research, 2003, 37(8): 1685-1690. doi: 10.1016/S0043-1354(02)00569-9 [5] BEN Y J, FU C X, HU M, et al. Human health risk assessment of antibiotic resistance associated with antibiotic residues in the environment: A review [J]. Environmental Research, 2019, 169: 483-493. doi: 10.1016/j.envres.2018.11.040 [6] ANIKA T, NOMAN Z, FERDOUS M, et al. Time dependent screening of antibiotic residues in milk of antibiotics treated cows [J]. Journal of Advanced Veterinary and Animal Research, 2019, 6(4): 516. doi: 10.5455/javar.2019.f376 [7] CHEN J, YING G G, DENG W J. Antibiotic residues in food: Extraction, analysis, and human health concerns [J]. Journal of Agricultural and Food Chemistry, 2019, 67(27): 7569-7586. doi: 10.1021/acs.jafc.9b01334 [8] BERGERON S, RAJ B, NATHANIEL R, et al. Presence of antibiotic resistance genes in raw source water of a drinking water treatment plant in a rural community of USA [J]. International Biodeterioration & Biodegradation, 2017, 124: 3-9. [9] CHEN C Q, LI J, CHEN P P, et al. Occurrence of antibiotics and antibiotic resistances in soils from wastewater irrigation areas in Beijing and Tianjin, China [J]. Environmental Pollution, 2014, 193: 94-101. doi: 10.1016/j.envpol.2014.06.005 [10] ASHBOLT N J, AMÉZQUITA A, BACKHAUS T, et al. Human Health Risk Assessment (HHRA) for environmental development and transfer of antibiotic resistance [J]. Environmental Health Perspectives, 2013, 121(9): 993-1001. doi: 10.1289/ehp.1206316 [11] CABANILLAS A G, CáCERES M I R, CAñAS M A Met al. Square wave adsorptive stripping voltametric determination of the mixture of nalidixic acid and its main metabolite (7-hydroxymethylnalidixic acid) by multivariate methods and artificial neural network [J]. Talanta, 2007, 72(3): 932-940. doi: 10.1016/j.talanta.2006.12.035 [12] RODRIGUEZ-NARVAEZ O M, PERALTA-HERNANDEZ J M, GOONETILLEKE A, et al. Treatment technologies for emerging contaminants in water: A review [J]. Chemical Engineering Journal, 2017, 323: 361-380. doi: 10.1016/j.cej.2017.04.106 [13] SUN Y Q, CHO D W, GRAHAM N J D, et al. Degradation of antibiotics by modified vacuum-UV based processes: Mechanistic consequences of H2O2 and K2S2O8 in the presence of halide ions [J]. Science of the Total Environment, 2019, 664: 312-321. doi: 10.1016/j.scitotenv.2019.02.006 [14] BENNEMLA M, CHABANI M, AMRANE A. Photocatalytic degradation of oxytetracycline in aqueous solutions with TiO2 in suspension and prediction by artificial neural networks [J]. International Journal of Chemical Kinetics, 2016, 48(8): 464-473. doi: 10.1002/kin.21005 [15] BAALOUDJ O, NASRALLAH N, KEBIR M, et al. Artificial neural network modeling of cefixime photodegradation by synthesized CoBi2O4 nanoparticles [J]. Environmental Science and Pollution Research, 2021, 28(12): 15436-15452. doi: 10.1007/s11356-020-11716-w [16] PHOON B L, ONG C C, MOHAMED SAHEED M S, et al. Conventional and emerging technologies for removal of antibiotics from wastewater [J]. Journal of Hazardous Materials, 2020, 400: 122961. doi: 10.1016/j.jhazmat.2020.122961 [17] LV H, TANG H Y. Machine learning methods and their application research[C]//2011 2nd International Symposium on Intelligence Information Processing and Trusted Computing. October 22-23, 2011, Wuhan, China. IEEE, 2011: 108-110. [18] WANG L D, ALEXANDER C A. Machine learning in big data [J]. International Journal of Mathematical, Engineering and Management Sciences, 2016, 1(2): 52-61. doi: 10.33889/IJMEMS.2016.1.2-006 [19] DASH S S, NAYAK S K, MISHRA D. A review on machine learning algorithms[M]//Smart Innovation, Systems and Technologies. Singapore: Springer Singapore, 2020: 495-507. [20] MA J, DING Y X, CHENG J C P, et al. A Lag-FLSTM deep learning network based on Bayesian Optimization for multi-sequential-variant PM2.5 prediction [J]. Sustainable Cities and Society, 2020, 60: 102237. doi: 10.1016/j.scs.2020.102237 [21] DELAVAR M, GHOLAMI A, SHIRAN G, et al. A novel method for improving air pollution prediction based on machine learning approaches: A case study applied to the capital city of Tehran [J]. ISPRS International Journal of Geo-Information, 2019, 8(2): 99. doi: 10.3390/ijgi8020099 [22] DIDELOT X, POUWELS K B. Machine-learning-assisted selection of antibiotic prescription [J]. Nature Medicine, 2019, 25(7): 1033-1034. doi: 10.1038/s41591-019-0517-0 [23] UYSAL CILOGLU F, SARIDAG A M, KILIC I H, et al. Identification of methicillin-resistant Staphylococcus aureus bacteria using surface-enhanced Raman spectroscopy and machine learning techniques [J]. The Analyst, 2020, 145(23): 7559-7570. doi: 10.1039/D0AN00476F [24] ZHONG Y S, NI Y N, KOKOT S. Application of differential pulse stripping voltammetry and chemometrics for the determination of three antibiotic drugs in food samples [J]. Chinese Chemical Letters, 2012, 23(3): 339-342. doi: 10.1016/j.cclet.2012.01.007 [25] ZENG K, WEI W, JIANG L, et al. Use of carbon nanotubes as a solid support to establish quantitative (centrifugation) and qualitative (filtration) immunoassays to detect gentamicin contamination in commercial milk [J]. Journal of Agricultural and Food Chemistry, 2016, 64(41): 7874-7881. doi: 10.1021/acs.jafc.6b03332 [26] YEHIA A M, ELBALKINY H T, RIAD S M, et al. Chemometrics for resolving spectral data of cephalosporines and tracing their residue in waste water samples [J]. Spectrochimica Acta Part A:Molecular and Biomolecular Spectroscopy, 2019, 219: 436-443. doi: 10.1016/j.saa.2019.04.081 [27] LONG Y, LI B, LIU H. Analysis of fluoroquinolones antibiotic residue in feed matrices using terahertz spectroscopy [J]. Applied Optics, 2018, 57(3): 544-550. doi: 10.1364/AO.57.000544 [28] URAPEN R, MASAWAT P. Novel method for the determination of tetracycline antibiotics in bovine milk based on digital-image-based colorimetry [J]. International Dairy Journal, 2015, 44: 1-5. doi: 10.1016/j.idairyj.2014.12.002 [29] ASADOLLAHI-BABOLI M, MANI-VARNOSFADERANI A. Rapid and simultaneous determination of tetracycline and cefixime antibiotics by mean of gold nanoparticles-screen printed gold electrode and chemometrics tools [J]. Measurement, 2014, 47: 145-149. doi: 10.1016/j.measurement.2013.08.029 [30] XU Z J, WANG Z K, LIU M Y, et al. Machine learning assisted dual-channel carbon quantum dots-based fluorescence sensor array for detection of tetracyclines [J]. Spectrochimica Acta Part A:Molecular and Biomolecular Spectroscopy, 2020, 232: 118147. doi: 10.1016/j.saa.2020.118147 [31] ATTIA K A M, NASSAR M W I, EL-ZEINY M B, et al. Effect of genetic algorithm as a variable selection method on different chemometric models applied for the analysis of binary mixture of amoxicillin and flucloxacillin: A comparative study [J]. Spectrochimica Acta. Part A, Molecular and Biomolecular Spectroscopy, 2016, 156: 54-62. doi: 10.1016/j.saa.2015.11.024 [32] DING W, ZHANG Y, KOU L P, et al. Electronic nose application for the determination of penicillin G in Saanen goat milk with fisher discriminate and multilayer perceptron neural network analyses [J]. Journal of Food Processing and Preservation, 2015, 39(6): 927-932. doi: 10.1111/jfpp.12305 [33] ZHAO L Z, DOU Y, GUO Y, et al. Artificial neural networks for non-destructive determination of acetylspiramycin powder by short-wavelength NIR spectroscopy [J]. Vibrational Spectroscopy, 2008, 47(1): 21-25. doi: 10.1016/j.vibspec.2008.01.007 [34] MORALES R, ORTIZ M C, SARABIA L A, et al. D-optimal designs and N-way techniques to determine sulfathiazole in milk by molecular fluorescence spectroscopy [J]. Analytica Chimica Acta, 2011, 707(1/2): 38-46. [35] MARTÍNEZ J L. Antibiotics and antibiotic resistance genes in natural environments [J]. Science, 2008, 321(5887): 365-367. doi: 10.1126/science.1159483 [36] SINGH R, SINGH A P, KUMAR S, et al. Antibiotic resistance in major rivers in the world: A systematic review on occurrence, emergence, and management strategies [J]. Journal of Cleaner Production, 2019, 234: 1484-1505. doi: 10.1016/j.jclepro.2019.06.243 [37] VELTRI D, KAMATH U, SHEHU A. Deep learning improves antimicrobial peptide recognition [J]. Bioinformatics, 2018, 34(16): 2740-2747. doi: 10.1093/bioinformatics/bty179 [38] 吴甜甜, 杨洁. 天然抗菌肽的研究进展及应用前景 [J]. 生物技术通报, 2009(1): 27-30. WU T T, YANG J. Progress in investigation of natural antimicrobial peptides and application prospect [J]. Biotechnology Bulletin, 2009(1): 27-30(in Chinese).
[39] MAGANA M, PUSHPANATHAN M, SANTOS A L, et al. The value of antimicrobial peptides in the age of resistance [J]. The Lancet Infectious Diseases, 2020, 20(9): e216-e230. doi: 10.1016/S1473-3099(20)30327-3 [40] POORINMOHAMMAD N, HAMEDI J, MOGHADDAM M H A M. Sequence-based analysis and prediction of lantibiotics: A machine learning approach [J]. Computational Biology and Chemistry, 2018, 77: 199-206. doi: 10.1016/j.compbiolchem.2018.10.004 [41] FJELL C D, JENSSEN H, HILPERT K, et al. Identification of novel antibacterial peptides by chemoinformatics and machine learning [J]. Journal of Medicinal Chemistry, 2009, 52(7): 2006-2015. doi: 10.1021/jm8015365 [42] SU X, XU J, YIN Y B, et al. Antimicrobial peptide identification using multi-scale convolutional network [J]. BMC Bioinformatics, 2019, 20(1): 730. doi: 10.1186/s12859-019-3327-y [43] LATA S, SHARMA B, RAGHAVA G. Analysis and prediction of antibacterial peptides [J]. BMC Bioinformatics, 2007, 8(1): 1-10. doi: 10.1186/1471-2105-8-1 [44] IRIYA R, JING W W, SYAL K, et al. Rapid antibiotic susceptibility testing based on bacterial motion patterns with long short- term memory neural networks [J]. IEEE Sensors Journal, 2020, 20(9): 4940-4950. doi: 10.1109/JSEN.2020.2967058 [45] GULL S, SHAMIM N, MINHAS F. AMAP: Hierarchical multi-label prediction of biologically active and antimicrobial peptides [J]. Computers in Biology and Medicine, 2019, 107: 172-181. doi: 10.1016/j.compbiomed.2019.02.018 [46] WHITE C, ISMAIL H D, SAIGO H, et al. CNN-BLPred: a Convolutional neural network based predictor for β-Lactamases (BL) and their classes [J]. BMC Bioinformatics, 2017, 18(16): 221-232. [47] HAMID M N, FRIEDBERG I. Identifying antimicrobial peptides using word embedding with deep recurrent neural networks [J]. Bioinformatics, 2019, 35(12): 2009-2016. doi: 10.1093/bioinformatics/bty937 [48] STOKES J M, YANG K, SWANSON K, et al. A deep learning approach to antibiotic discovery [J]. Cell, 2020, 180(4): 688-702.e13. doi: 10.1016/j.cell.2020.01.021 [49] ZOFFMANN S, VERCRUYSSE M, BENMANSOUR F, et al. Machine learning-powered antibiotics phenotypic drug discovery [J]. Scientific Reports, 2019, 9: 5013. doi: 10.1038/s41598-019-39387-9 [50] KOHANSKI M A, DWYER D J, COLLINS J J. How antibiotics kill bacteria: From targets to networks [J]. Nature Reviews Microbiology, 2010, 8(6): 423-435. doi: 10.1038/nrmicro2333 [51] LEE M W, de ANDA J, KROLL C, et al. How do cyclic antibiotics with activity against Gram-negative bacteria permeate membranes?A machine learning informed experimental study [J]. Biochimica et Biophysica Acta (BBA) - Biomembranes, 2020, 1862(8): 183302. doi: 10.1016/j.bbamem.2020.183302 [52] LU X Q, WANG L, de LIU H, et al. Studies on the interaction between antibiotics and DNA [J]. Talanta, 2007, 73(3): 444-450. doi: 10.1016/j.talanta.2007.04.006 [53] BROWN C, TSENG D, LARKIN P M K, et al. Automated, cost-effective optical system for accelerated antimicrobial susceptibility testing (AST) using deep learning [J]. ACS Photonics, 2020, 7(9): 2527-2538. doi: 10.1021/acsphotonics.0c00841 [54] YANG J H, WRIGHT S N, HAMBLIN M, et al. A white-box machine learning approach for revealing antibiotic mechanisms of action [J]. Cell, 2019, 177(6): 1649-1661.e9. doi: 10.1016/j.cell.2019.04.016 [55] NIELSEN J. Antibiotic lethality is impacted by nutrient availabilities: New insights from machine learning [J]. Cell, 2019, 177(6): 1373-1374. doi: 10.1016/j.cell.2019.05.015 [56] BUDAK F, ÜBEYLI E D. Detection of resistivity for antibiotics by probabilistic neural networks [J]. Journal of Medical Systems, 2011, 35(1): 87-91. doi: 10.1007/s10916-009-9344-z [57] BINH V N, DANG N, ANH N T K, et al. Antibiotics in the aquatic environment of Vietnam: Sources, concentrations, risk and control strategy [J]. Chemosphere, 2018, 197: 438-450. doi: 10.1016/j.chemosphere.2018.01.061 [58] WANG Q, ZHAO W M. Optical methods of antibiotic residues detections: A comprehensive review [J]. Sensors and Actuators B:Chemical, 2018, 269: 238-256. [59] ZHOU J W, XU Z H, CHEN S W. Simulation and prediction of the thuringiensin abiotic degradation processes in aqueous solution by a radius basis function neural network model [J]. Chemosphere, 2013, 91(4): 442-447. doi: 10.1016/j.chemosphere.2012.11.062 [60] ZHU X Z, WAN Z H, TSANG D C W, et al. Machine learning for the selection of carbon-based materials for tetracycline and sulfamethoxazole adsorption [J]. Chemical Engineering Journal, 2021, 406: 126782. doi: 10.1016/j.cej.2020.126782 [61] FOROUGHI M, AHMADI AZQHANDI M H, KAKHKI S. Bio-inspired, high, and fast adsorption of tetracycline from aqueous media using Fe3O4-g-CN@PEI-β-CD nanocomposite: Modeling by response surface methodology (RSM), boosted regression tree (BRT), and general regression neural network (GRNN) [J]. Journal of Hazardous Materials, 2020, 388: 121769. doi: 10.1016/j.jhazmat.2019.121769 [62] TABATABAI-YAZDI F S, EBRAHIMIAN PIRBAZARI A, ESMAEILI KHALILSARAEI F, et al. Photocatalytic treatment of tetracycline antibiotic wastewater by silver/TiO 2 nanosheets/reduced graphene oxide and artificial neural network modeling [J]. Water Environment Research, 2020, 92(5): 662-676. doi: 10.1002/wer.1258 [63] TALWAR S, VERMA A K, SANGAL V K. Modeling and optimization of fixed mode dual effect (photocatalysis and photo-Fenton) assisted Metronidazole degradation using ANN coupled with genetic algorithm [J]. Journal of Environmental Management, 2019, 250: 109428. doi: 10.1016/j.jenvman.2019.109428 [64] DONUT N, CAVAS L. Artificial neural network modeling of tetracycline biosorption by pre-treated Posidonia oceanic [J]. Turkish Journal of Fisheries and Aquatic Sciences, 2017, 17(6): 1317-1333. [65] ALAVI N, SARMADI K, GOUDARZI G, et al. Attenuation of tetracyclines during chicken manure and bagasse co-composting: Degradation, kinetics, and artificial neural network modeling [J]. Journal of Environmental Management, 2019, 231: 1203-1210. doi: 10.1016/j.jenvman.2018.11.003 [66] JAMALI ALYANI S, EBRAHIMIAN PIRBAZARI A, ESMAEILI KHALILSARAEI F, et al. Growing Co-doped TiO2 nanosheets on reduced graphene oxide for efficient photocatalytic removal of tetracycline antibiotic from aqueous solution and modeling the process by artificial neural network [J]. Journal of Alloys and Compounds, 2019, 799: 169-182. doi: 10.1016/j.jallcom.2019.05.175 [67] TALWAR S, SANGAL V K, VERMA A, et al. Modeling, optimization and kinetic study for photocatalytic treatment of ornidazole using slurry and fixed-bed approach [J]. Arabian Journal for Science and Engineering, 2018, 43(11): 6191-6202. doi: 10.1007/s13369-018-3388-7 [68] YURTSEVER U, CAN DOĞAN E, GENÇ N. The use of output-dependent data scaling with artificial neural networks and multilinear regression for modeling of ciprofloxacin removal from aqueous solution [J]. Journal of Water Reuse and Desalination, 2017, 7(1): 25-36. doi: 10.2166/wrd.2016.099 [69] OLADIPO A A, ABUREESH M A, GAZI M. Bifunctional composite from spent “Cyprus coffee” for tetracycline removal and phenol degradation: Solar-Fenton process and artificial neural network [J]. International Journal of Biological Macromolecules, 2016, 90: 89-99. doi: 10.1016/j.ijbiomac.2015.08.054 [70] PELALAK R, ALIZADEH R, GHARESHABANI E, et al. Degradation of sulfonamide antibiotics using ozone-based advanced oxidation process: Experimental, modeling, transformation mechanism and DFT study [J]. Science of the Total Environment, 2020, 734: 139446. doi: 10.1016/j.scitotenv.2020.139446 [71] SAMADI-MAYBODI A, NIKOU M. Removal of sarafloxacin from aqueous solution by a magnetized metal-organic framework;Artificial neural network modeling [J]. Polyhedron, 2020, 179: 114342. doi: 10.1016/j.poly.2019.114342 [72] CHOWDHURY S, HALDER G, MANDAL T, et al. Cetylpyridinium bromide assisted micellar-enhanced ultrafiltration for treating enrofloxacin-laden water [J]. Science of the Total Environment, 2019, 687: 10-23. doi: 10.1016/j.scitotenv.2019.06.074 [73] CHOWDHURY S, SIKDER J, MANDAL T, et al. Comprehensive analysis on sorptive uptake of enrofloxacin by activated carbon derived from industrial paper sludge [J]. Science of the Total Environment, 2019, 665: 438-452. doi: 10.1016/j.scitotenv.2019.02.081 [74] YABALAK E. Degradation of ticarcillin by subcritical water oxidation method: Application of response surface methodology and artificial neural network modeling [J]. Journal of Environmental Science and Health, Part A, 2018, 53(11): 975-985. doi: 10.1080/10934529.2018.1471023 [75] SINGH K P, SINGH A K, GUPTA S, et al. Modeling and optimization of reductive degradation of chloramphenicol in aqueous solution by zero-valent bimetallic nanoparticles [J]. Environmental Science and Pollution Research, 2012, 19(6): 2063-2078. doi: 10.1007/s11356-011-0700-4 [76] HUANG B, WANG H C, CUI D, et al. Treatment of pharmaceutical wastewater containing β-lactams antibiotics by a pilot-scale anaerobic membrane bioreactor (AnMBR) [J]. Chemical Engineering Journal, 2018, 341: 238-247. doi: 10.1016/j.cej.2018.01.149 [77] ZHOU C Z, WANG Q, ZHOU C Y. Photocatalytic degradation of antibiotics by molecular assembly porous carbon nitride: Activity studies and artificial neural networks modeling [J]. Chemical Physics Letters, 2020, 750: 137479. doi: 10.1016/j.cplett.2020.137479 [78] HU D X, MIN H C, CHEN Z B, et al. Performance improvement and model of a bio-electrochemical system built-in up-flow anaerobic sludge blanket for treating β-lactams pharmaceutical wastewater under different hydraulic retention time [J]. Water Research, 2019, 164: 114915. doi: 10.1016/j.watres.2019.114915