基于深度学习的金属离子-有机配体配位稳定常数的预测

doi:10.11949/0438-1157.20221081

摘要/Abstract

摘要：

金属离子-有机配体配合物的稳定性由金属离子种类、有机配体结构和实验条件三个因素决定。使用传统方法获得配合物的稳定常数耗时耗力，不利于特异性金属螯合剂的高通量筛选。因此，基于多头图注意力网络提出一种综合考虑多种影响因素的配合物稳定常数高通量预测模型。首先，对从数据库中提取的7127个配合物中涉及到的1371个有机分子生成分子属性图，其次利用多头图注意力网络对分子属性图进行特征提取，并将提取的分子特征拼接编码的金属离子和实验条件，最后送入全连接层进行配合物稳定常数的预测。模型在测试集上的R²和RMSE分别为0.956和1.251，表明所构建模型具有良好的泛化能力。此外，将文献中基于密度泛函理论计算的螯合物稳定常数与模型预测值进行对比，结果表明模型更为可靠与高效。

关键词: 配合物, 稳定常数, 多头图注意力网络, 深度学习

Abstract:

The stability of metal ion-organic ligand complexes is determined by three factors: metal ion species, organic ligand structure and experimental conditions. Obtaining the stability constants of complexes using traditional methods is time-consuming and labor-intensive, which is not conducive to high-throughput screening of specific metal chelators. Therefore, based on multi-head graph attention network (multi-head GAT), a high-throughput prediction model of complex stability constant is proposed in this paper. Firstly, molecular attribute diagrams were generated for 1371 organic molecules out of 7127 complexes extracted from mini stability constant database. Second, the multi-head graph attention network is used to extract the features of the attributed molecular graph. The extracted molecular features are spliced with the metal ions and the experimental conditions encoded by one-hot encoding. Finally, all feature codes are sent to the fully connected layer to predict the stability constants of the complexes. The determination coefficient (R²) and root mean square error (RMSE) of the model on the test set are 0.956 and 1.251, respectively, indicating that the model has good generalization ability. In addition, using the model to predict the stability constants of chelates in the literature, the model proposed in this paper is more reliable and efficient than the results based on density functional theory (DFT) calculations.

Key words: complex, stability constant, multi-head graph attention network, deep learning

中图分类号:

TQ 317.5

齐书平, 王文龙, 张磊, 都健. 基于深度学习的金属离子-有机配体配位稳定常数的预测[J]. 化工学报, 2022, 73(12): 5461-5468.

Shuping QI, Wenlong WANG, Lei ZHANG, Jian DU. A deep learning-based model for predicting the stability constants of metal ions with organic ligands[J]. CIESC Journal, 2022, 73(12): 5461-5468.

图/表 5

参考文献 30

1	Lawrance G A. Introduction to Coordination Chemistry[M]. Chichester, UK: John Wiley & Sons, 2010.
2	Martell A E, Motekaitis R J, Chen D, et al. Selection of new Fe(Ⅲ)/Fe(Ⅱ) chelating agents as catalysts for the oxidation of hydrogen sulfide to sulfur by air[J]. Canadian Journal of Chemistry, 1996, 74(10): 1872-1879.
3	Kuźnik N, Wyskocka M. Iron(Ⅲ) contrast agent candidates for MRI: a survey of the structure-effect relationship in the last 15 years of studies[J]. European Journal of Inorganic Chemistry, 2016, 2016(4): 445-458.
4	Checa-Fernandez A, Santos A, Romero A, et al. Application of chelating agents to enhance Fenton process in soil remediation: a review[J]. Catalysts, 2021, 11(6): 722.
5	Wilson A M, Bailey P J, Tasker P A, et al. Solvent extraction: the coordination chemistry behind extractive metallurgy[J]. Chemical Society Reviews, 2014, 43(1): 123-134.
6	Gaur K, Pérez Otero S C, Benjamín-Rivera J A, et al. Iron chelator transmetalative approach to inhibit human ribonucleotide reductase[J]. Journal of the American Chemistry Society Au, 2021, 1(6): 865-878.
7	Sheikh T A, Arshad M N, Rahman M M, et al. Trace electrochemical detection of Ni²⁺ ions with bidentate N,N'(ethane-1,2-diyl)bis(3,4-dimethoxybenzenesulfonamide)[EDBDMBS] as a chelating agent[J]. Inorganica Chimica Acta, 2017, 464: 157-166.
8	Solov'ev V, Marcou G, Tsivadze A, et al. Complexation of Mn²⁺, Fe²⁺, Y³⁺, La³⁺, Pb²⁺, and UO₂ ²⁺ with organic ligands: QSPR ensemble modeling of stability constants[J]. Industrial & Engineering Chemistry Research, 2012, 51(41): 13482-13489.
9	Solov'ev V P, Varnek A, Wipff G. Modeling of ion complexation and extraction using substructural molecular fragments[J]. Journal of Chemical Information and Computer Sciences, 2000, 40(3): 847-858.
10	Solov'ev V, Sukhno I, Buzko V, et al. Stability constants of complexes of Zn²⁺, Cd²⁺, and Hg²⁺ with organic ligands: QSPR consensus modeling and design of new metal binders[J]. Journal of Inclusion Phenomena and Macrocyclic Chemistry, 2012, 72(3/4): 309-321.
11	Solov'ev V P, Tsivadze A Y, Varnek A A. New approach for accurate QSPR modeling of metal complexation: application to stability constants of complexes of lanthanide ions Ln³⁺, Ag⁺, Zn²⁺, Cd²⁺ and Hg²⁺ with organic ligands in water[J]. Macroheterocycles, 2012, 5(4/5): 404-410.
12	Solov'ev V P, Kireeva N, Tsivadze A Y, et al. QSPR ensemble modelling of alkaline-earth metal complexation[J]. Journal of Inclusion Phenomena and Macrocyclic Chemistry, 2013, 76(1/2): 159-171.
13	Varnek A, Solov‚Äôev V. Quantitative structure-property relationships in solvent extraction and complexation of metals[M]//Ion Exchange and Solvent Extraction Series. Florida, USA: CRC Press, 2009: 319-358.
14	Solov'ev V, Varnek A, Tsivadze A. QSPR ensemble modelling of the 1∶1 and 1∶2 complexation of Co²⁺, Ni²⁺, and Cu²⁺ with organic ligands: relationships between stability constants[J]. Journal of Computer-Aided Molecular Design, 2014, 28(5): 549-564.
15	Solov'ev V, Kireeva N, Ovchinnikova S, et al. The complexation of metal ions with various organic ligands in water: prediction of stability constants by QSPR ensemble modelling[J]. Journal of Inclusion Phenomena and Macrocyclic Chemistry, 2015, 83(1/2): 89-101.
16	Mitrofanov A A, Matveev P I, Yakubova K V, et al. Deep learning insights into lanthanides complexation chemistry[J]. Molecules (Basel, Switzerland), 2021, 26(11): 3237.
17	肖琰, 曹晨忠. 过渡金属-二芳基希夫碱配位稳定常数的QSPR研究[J]. 化学研究与应用, 2020, 32(5): 782-788.
	Xiao Y, Cao C Z. QSPR study of the stability constant of the complexes involving bi-aryl Schiff bases reacting with transition metals[J]. Chemical Research and Application, 2020, 32(5): 782-788.
18	Chaube S, Srinivasan S G, Rai B. Applied machine learning for predicting the lanthanide-ligand binding affinities[J]. Scientific Reports, 2020, 10: 14322.
19	Velickovic P, Cucurull G, Casanova A, et al. Graph attention networks[EB/OL]. 2018. .
20	梁春余, 杨家振, 刘祁涛. 支持电解质和离子强度对溶液中配合物稳定性的影响(Ⅲ): Pitzer理论在确定配合物热力学稳定常数中的应用[J]. 化学学报, 1986, 44(3): 213-219.
	Liang C Y, Yang J Z, Liu Q T. The influence of ionic strength and supporting electrolytes on the stability of complexes in solutions(Ⅲ): The application of Pitzer's equation to determine the thermodynamic stability constant of complexes[J]. Acta Chimica Sinica, 1986, 44(3): 213-219.
21	Pettit L. The IUPAC stability constants database[J]. Chemistry International, 2006, 28: 14-15.
22	Wang M Y. Deep graph library: towards efficient and scalable deep learning on graphs[C]//ICLR Workshop on Representation Learning on Graphs and Manifolds. ICLR, 2019.
23	RDKit: open-source cheminformatics software[CP/OL]. [2020-10-26]. .
24	Coley C W, Barzilay R, Green W H, et al. Convolutional embedding of attributed molecular graphs for physical property prediction[J]. Journal of Chemical Information and Modeling, 2017, 57(8): 1757-1772.
25	Imambi S, Prakash K B, Kanagachidambaresan G R. PyTorch[M]//Programming with TensorFlow. Cham: Springer International Publishing, 2021: 87-104.
26	Abadi M, Barham P, Chen J, et al. TensorFlow: a system for large-scale machine learning[C]//12th USENIX Symposium on Operating Systems Design and Implementation (OSDI 16). Atlanta, Georgia, USA: USENIX Association, 2016.
27	Mechachti F, Lakehal S, Lakehal A, et al. Predicted structure and selectivity of 3d transition metal complexes with glutamic N, N-bis(carboxymethyl) acid[J]. New Journal of Chemistry, 2021, 45(39): 18366-18378.
28	Chauhan G, Stein M, Seidel-Morgenstern A, et al. The thermodynamics and biodegradability of chelating agents upon metal extraction[J]. Chemical Engineering Science, 2015, 137: 768-785.
29	Ridzwan M H, Yaakob M K, Zabidi Z M, et al. Computational insight into the quantum chemistry, interaction and adsorption energy of aminopolycarboxylic acid chelating agents towards metal cations[J]. Computational and Theoretical Chemistry, 2022, 1208: 113579.
30	Cassone G, Chillè D, Giacobello F, et al. Interaction between AS(Ⅲ) and simple thioacids in water: an experimental and ab initio molecular dynamics investigation[J]. The Journal of Physical Chemistry B, 2019, 123(28): 6090-6098.

序号	原子特征	编码	序号	原子特征	编码	序号	键特征	编码
0~11	原子序数^①	独热	37	是否在环上^①	独热	44~47	键类型^①	数值
12~15	总氢数^①	独热	38	Gasteiger电荷^②	数值	48	是否共轭^①	数值
16~22	显式化合价^①	独热	39	Gasteiger氢电荷^②	数值	49~50	键方向^①	数值
23~25	形式电荷^①	独热	40	Crippen摩尔折射率^②	数值	51~52	键构型^①	数值
26~32	杂化方式^①	独热	41	Crippen logP^②	数值	53	立体键^①	数值
33~35	手性^①	独热	42	TPSA^②	数值
36	芳香性^①	独热	43	LabuteASAContribs^②	数值

序号	原子特征	编码	序号	原子特征	编码	序号	键特征	编码
0~11	原子序数^①	独热	37	是否在环上^①	独热	44~47	键类型^①	数值
12~15	总氢数^①	独热	38	Gasteiger电荷^②	数值	48	是否共轭^①	数值
16~22	显式化合价^①	独热	39	Gasteiger氢电荷^②	数值	49~50	键方向^①	数值
23~25	形式电荷^①	独热	40	Crippen摩尔折射率^②	数值	51~52	键构型^①	数值
26~32	杂化方式^①	独热	41	Crippen logP^②	数值	53	立体键^①	数值
33~35	手性^①	独热	42	TPSA^②	数值
36	芳香性^①	独热	43	LabuteASAContribs^②	数值

序号	金属离子	配位态	lgK预测值	lgK DFT计算值	lgK实验值	绝对误差
1^[27]	Mn²⁺	单配位	7.52	48.65	7.60	0.08
	Fe²⁺	单配位	8.64	51.95	8.70	0.06
	Co²⁺	单配位	9.52	55.19	10.00	0.48
	Ni²⁺	单配位	9.80	63.96	10.90	1.10
2^[28]	Ni²⁺	单配位	12.10	7.94	11.51	0.59
3^[28]	Ni²⁺	单配位	13.24	16.78	17.95	4.71
	Co²⁺	单配位	13.65	14.05	15.41	1.76

序号	金属离子	配位态	lgK预测值	lgK DFT计算值	lgK实验值	绝对误差
1^[27]	Mn²⁺	单配位	7.52	48.65	7.60	0.08
	Fe²⁺	单配位	8.64	51.95	8.70	0.06
	Co²⁺	单配位	9.52	55.19	10.00	0.48
	Ni²⁺	单配位	9.80	63.96	10.90	1.10
2^[28]	Ni²⁺	单配位	12.10	7.94	11.51	0.59
3^[28]	Ni²⁺	单配位	13.24	16.78	17.95	4.71
	Co²⁺	单配位	13.65	14.05	15.41	1.76

[1]	温凯杰, 郭力, 夏诏杰, 陈建华. 一种耦合CFD与深度学习的气固快速模拟方法[J]. 化工学报, 2023, 74(9): 3775-3785.
[2]	张红锐, 张田, 隆曦孜, 李先宁. 光催化与微生物燃料电池耦合对Cu-EDTA的降解特性[J]. 化工学报, 2022, 73(5): 2149-2157.
[3]	阳源源, 王进芝, 杜俊哲, 杜奥冰, 赵井文, 崔光磊. 镁-氯溶剂化结构在镁基电解液中的作用[J]. 化工学报, 2021, 72(6): 3116-3129.
[4]	谢昊源, 黄群星, 林晓青, 李晓东, 严建华. 基于图像深度学习的垃圾热值预测研究[J]. 化工学报, 2021, 72(5): 2773-2782.
[5]	陈忠圣, 朱梅玉, 贺彦林, 徐圆, 朱群雄. 基于分位数回归CGAN的虚拟样本生成方法及其过程建模应用[J]. 化工学报, 2021, 72(3): 1529-1538.
[6]	于程远, 吴金奎, 周利, 吉旭, 戴一阳, 党亚固. 基于深度学习预测有机光伏电池能量转换效率[J]. 化工学报, 2021, 72(3): 1487-1495.
[7]	王晓慧, 王延江, 邓晓刚, 张政. 基于加权深度支持向量数据描述的工业过程故障检测[J]. 化工学报, 2021, 72(11): 5707-5716.
[8]	尹林子, 关羽吟, 蒋朝辉, 许雪梅. 基于k-means++的高炉铁水硅含量数据优选方法[J]. 化工学报, 2020, 71(8): 3661-3670.
[9]	孙瑞阳, 陆颖舟, 孟洪, 李春喜. 噻吩硫化物的硝化产物用于模型油中DBT的高效吸附[J]. 化工学报, 2020, 71(5): 2256-2264.
[10]	田璐瑶, 王梓豪, 粟杨, 文华强, 申威峰. 基于深度学习的溶剂定量构效关系建模研究进展[J]. 化工学报, 2020, 71(10): 4462-4472.
[11]	黄正梁, 王超, 李少硕, 杨遥, 孙婧元, 王靖岱, 阳永荣. 基于深度学习的气液固三相反应器图像分析方法及应用[J]. 化工学报, 2020, 71(1): 274-282.
[12]	朱啸林, 徐存英, 唐杰, 华一新, 张启波, 刘海, 王祥, 黄梦婷. 氯化铬和氯化镍浓度及其存在形式对ChCl-EG低共熔溶剂电导率的影响[J]. 化工学报, 2019, 70(7): 2786-2794.
[13]	张鹏, 王雨露, 丁文杰, 赵文麟. 新型巯基絮凝剂PAM-GSH的合成及除Mn(Ⅱ)性能研究[J]. 化工学报, 2019, 70(5): 1932-1941.
[14]	顾恒昌, 牟鹏, 李建伟. 基于交叉迭代BLSTM网络的乙烯裂解炉建模[J]. 化工学报, 2019, 70(2): 548-555.
[15]	董顺, 李益国, 孙栓柱, 刘西陲, 沈炯. 基于状态空间主成分分析网络的故障检测方法[J]. 化工学报, 2018, 69(8): 3528-3536.