基于层级和全局特征结合的蛋白质序列EC编号预测

doi:10.16288/j.yczz.24-102

遗传 ›› 2024, Vol. 46 ›› Issue (8): 661-669.doi: 10.16288/j.yczz.24-102

• 技术与方法 • 上一篇

基于层级和全局特征结合的蛋白质序列EC编号预测

杨帆¹^,²^,³^,⁴(), 韩巧玲¹^,²^,³^,⁴, 赵文迪¹^,²^,³^,⁴, 赵玥¹^,²^,³^,⁴()

1.北京林业大学工学院，北京 100083
2.林业装备与自动化国家林业局重点实验室，北京 100083
3.城乡生态环境北京实验室，北京 100083
4.北京林业大学智慧林业研究中心，北京 100083

收稿日期:2024-04-12 修回日期:2024-06-18 出版日期:2024-08-20 发布日期:2024-06-19
通讯作者: 赵玥，博士，教授，研究方向：人工智能、图像处理和模式识别等。E-mail: zhaoyue0609@126.com
作者简介:杨帆，硕士研究生，专业方向：机器学习与数据处理。E-mail: yangfan_muyi@163.com
基金资助:
国家自然科学基金面上项目(32071838);国家自然科学基金青年科学基金项目(32101590)

EC number prediction of protein sequences based on combination of hierarchical and global features

Fan Yang¹^,²^,³^,⁴(), Qiaoling Han¹^,²^,³^,⁴, Wendi Zhao¹^,²^,³^,⁴, Yue Zhao¹^,²^,³^,⁴()

1. School of technology, Beijing Forestry University, Beijing 100083, China
2. Key Lab of State Forestry Administration for Forestry Equipment and Automation, Beijing 100083, China
3. Beijing Laboratory of Urban and Rural Ecological Environment, Beijing 100083, China
4. Research Center for Intelligent Forestry, Beijing Forestry University, Beijing 100083, China

Received:2024-04-12 Revised:2024-06-18 Published:2024-08-20 Online:2024-06-19
Supported by:
National Natural Science Foundation of China(32071838);National Natural Science Youth Foundation of China(32101590)

摘要/Abstract

摘要：

酶功能的识别对理解生命活动的机制、推进生命科学的发展有重要作用。然而现有的酶EC编号预测方法，并未充分利用蛋白质序列信息，在识别精度上仍有所不足。针对上述问题，本研究提出一种基于层级特征和全局特征的EC编号预测网络(EC number prediction network using hierarchical features and global features, ECPN-HFGF)。该方法首先通过残差网络提取蛋白质序列通用特征，并通过层级特征提取模块和全局特征提取模块进一步提取蛋白质序列的层级特征和全局特征，之后结合两种特征信息的预测结果，采用多任务学习框架，实现酶EC编号的精确预测。计算实验结果表明，ECPN-HFGF方法在蛋白质序列EC编号预测任务上性能最佳，宏观F1值和微观F1值分别达到95.5%和99.0%。ECPN-HFGF方法能有效结合蛋白质序列的层级特征和全局特征，快速准确预测蛋白质序列EC编号，比当前常用方法预测精确度更高，能够为酶学研究和酶工程应用的发展提供一种高效的思路和方法。

关键词: 酶功能预测, 蛋白质序列, 深度学习, 层级多标签分类, 全局特征

Abstract:

The identification of enzyme functions plays a crucial role in understanding the mechanisms of biological activities and advancing the development of life sciences. However, existing enzyme EC number prediction methods did not fully utilize protein sequence information and still had shortcomings in identification accuracy. To address this issue, we proposed an EC number prediction network using hierarchical features and global features (ECPN-HFGF). This method first utilized residual networks to extract generic features from protein sequences, and then employed hierarchical feature extraction modules and global feature extraction modules to further extract hierarchical and global features of protein sequences. Subsequently, the prediction results of both feature types were combined, and a multitask learning framework was utilized to achieve accurate prediction of enzyme EC numbers. Experimental results indicated that the ECPN-HFGF method performed best in the task of predicting EC numbers for protein sequences, achieving macro F1 and micro F1 scores of 95.5% and 99.0%, respectively. The ECPN-HFGF method effectively combined hierarchical and global features of protein sequences, allowing for rapid and accurate EC number prediction. Compared to current commonly used methods, this method offers significantly higher prediction accuracy, providing an efficient approach for the advancement of enzymology research and enzyme engineering applications.

Key words: enzyme function prediction, protein sequence, deep learning, hierarchical multi-label classification, global feature

杨帆, 韩巧玲, 赵文迪, 赵玥. 基于层级和全局特征结合的蛋白质序列EC编号预测[J]. 遗传, 2024, 46(8): 661-669.

Fan Yang, Qiaoling Han, Wendi Zhao, Yue Zhao. EC number prediction of protein sequences based on combination of hierarchical and global features[J]. Hereditas(Beijing), 2024, 46(8): 661-669.

图/表 8

表1

图1

图2

图3

图4

表2

表3

图5

参考文献 35

[1]	Aaron C. The ubiquitin-proteasome pathway: on protein death and cell life. EMBO J, 1998, 17(24): 7151-7160. doi: 10.1093/emboj/17.24.7151 pmid: 9857172
[2]	Satam H, Joshi K, Mangrolia U, Waghoo S, Zaidi G, Rawool S, Thakare RP, Banday S, Mishra AK, Das G, Malonia SK. Next-generation sequencing technology: current trends and advancements. Biology (Basel), 2023, 12(7): 997.
[3]	UniProt Consortium. UniProt: the universal protein knowledgebase in 2023. Nucleic Acids Res, 2023, 51(D1): D523-D531.
[4]	Yan TC, Yue ZX, Xu HQ, Liu YH, Hong YF, Chen GX, Tao L, Xie T. A systematic review of state-of-the-art strategies for machine learning-based protein function prediction. Comput Biol Med, 2023, 154: 106446.
[5]	McDonald AG, Tipton KF. Enzyme nomenclature and classification: the state of the art. FEBS J, 2023, 290(9): 2214-2231.
[6]	Markus B, C GC, Andreas K, Arkadij K, Stefan L, Gustav O, Elina S, Radka S. Accelerating biocatalysis discovery with machine learning: a paradigm shift in enzyme engineering, discovery, and design. ACS Catal, 2023, 13(21): 14454-14469. doi: 10.1021/acscatal.3c03417 pmid: 37942268
[7]	Ai YJ, He MQ, Liang QL. Recent research progress in nanozyme and their biomedical applications. Chin Pharm J, 2023, 58(20): 1801-1813. doi: 10.11669/cpj.2023.20.001
	艾永建, 何梦崎, 梁琼麟. 纳米酶在生物医药领域应用的研究进展. 中国药学杂志, 2023, 58(20): 1801-1813. doi: 10.11669/cpj.2023.20.001
[8]	Du WZ, Li YJ, Wu JL, Chen SY, Jiang L, Liu G, Xie N. Identification and functional study of AA11 family polysaccharide monooxygenase genes in filamentous fungus Podospora anserina. Hereditas(Beijing), 2023, 45(12): 1128-1146.
	杜文珍, 李元敬, 吴佳玲, 陈思羽, 姜亮, 刘刚, 谢宁. 丝状真菌Podospora anserina AA11家族裂解多糖单加氧酶基因的鉴定和功能研究. 遗传, 2023, 45(12): 1128-1146.
[9]	Shang XK, Zhang SM, Ni JJ. Research progress of cathepsin B in brain aging and Alzheimer’s diseases. Hereditas(Beijing), 2023, 45(3): 212-220.
	商晓康, 张思萌, 倪军军. 组织蛋白酶B参与脑衰老及阿尔兹海默症发生发展研究进展. 遗传, 2023, 45(3): 212-220.
[10]	Wang CW, Shui K, Ma SS, Lin SF, Zhang Y, Wen B, Deng WK, Xu HD, Hu H, Guo AY, Xue Y, Zhang LY. Integrated omics in Drosophila uncover a circadian kinome. Nat Commun, 2020, 11(1): 2710.
[11]	Wang M, Wang LY, Liu HZ, Chen JJ, Liu D. Transcriptome analyses implicate endogenous retroviruses involved in the host antiviral immune system through the interferon pathway. Virol Sin, 2021, 36(6):1315-1326. doi: 10.1007/s12250-021-00370-2 pmid: 34009516
[12]	Li SJ, Zhao J, Xiao R. A literature review of the application of enzymes in processing of plant-based foods for elderly people. Food Science, 2019, 40(21): 350-356. doi: 10.7506/spkx1002-6630-20181011-089
	李双娇, 赵静, 肖蓉. 酶在植物源性老年食品加工中的应用研究进展. 食品科学, 2019, 40(21): 350-356. doi: 10.7506/spkx1002-6630-20181011-089
[13]	Victorino da Silva Amatto I, Gonsales da Rosa-Garzon N, Antônio de Oliveira Simões F, Santiago F, Pereira da Silva Leite N, Raspante Martins J, Cabral H. Enzyme engineering and its industrial applications. Biotechnol Appl Biochem, 2022, 69(2): 389-409.
[14]	East KW, Skeens E, Cui JY, Belato HB, Mitchell B, Hsu R, Batista VS, Palermo G, Lisi GP. NMR and computational methods for molecular resolution of allosteric pathways in enzyme complexes. Biophys Rev, 2020, 12(1): 155-174. doi: 10.1007/s12551-019-00609-z pmid: 31838649
[15]	Vedel IM, Papagiannoula A, Naudi-Fabra S, Milles S. Nuclear magnetic resonance/single molecule fluorescence combinations to study dynamic protein systems. Curr Opin Struct Biol, 2023, 82: 102659.
[16]	Yabukarski F, Biel JT, Pinney MM, Doukov T, Powers AS, Fraser JS, Herschlag D. Assessment of enzyme active site positioning and tests of catalytic mechanisms through X-ray-derived conformational ensembles. Proc Natl Acad Sci USA, 2020, 117(52): 33204-33215. doi: 10.1073/pnas.2011350117 pmid: 33376217
[17]	Dalkiran A, Rifaioglu AS, Martin MJ, Cetin-Atalay R, Atalay V, Doğan T. ECPred: a tool for the prediction of the enzymatic functions of protein sequences based on the EC nomenclature. BMC Bioinformatics, 2018, 19(1): 334. doi: 10.1186/s12859-018-2368-y pmid: 30241466
[18]	Chi FY, Li C, Feng XD. Advances in machine learning for predicting protein functions. Chin J Biotechnol, 2023, 39(6): 2141-2157.
	池燕飞, 李春, 冯旭东. 机器学习在蛋白质功能预测领域的研究进展. 生物工程学报, 2023, 39(6): 2141-2157.
[19]	Chen MM, Zhang WZ, Gou YJ, Xu DY, Wei YX, Liu D, Han C, Huang XH, Li CZ, Ning WS, Peng D, Xue Y. GPS 6.0: an updated server for prediction of kinase-specific phosphorylation sites in proteins. Nucleic Acids Res, 2023, 51(W1): W243-W250. doi: 10.1093/nar/gkad383 pmid: 37158278
[20]	Ruan CF, Ning WS, Liu Z, Zhang XL, Fang Z, Li YN, Dang YJ, Xue Y, Ye ML. Precipitate-supported thermal proteome profiling coupled with deep learning for comprehensive screening of drug target proteins. ACS Chem Biol, 2022, 17(1): 252-262. doi: 10.1021/acschembio.1c00936 pmid: 34989232
[21]	Shui K, Wang CW, Zhang XD, Ma SS, Li QY, Ning WS, Zhang WZ, Chen MM, Peng D, Hu H, Fang Z, Guo AY, Gao GJ, Ye ML, Zhang LY, Xue Y. Small-sample learning reveals propionylation in determining global protein homeostasis. Nat Commun, 2023, 14(1): 2813. doi: 10.1038/s41467-023-38414-8 pmid: 37198164
[22]	Zaru R, Orchard S; UniProt Consortium. UniProt tools: bLAST, align, peptide search, and ID mapping. Curr Protoc, 2023, 3(3): e697.
[23]	Edgar RC. Muscle5: high-accuracy alignment ensembles enable unbiased assessments of sequence homology and phylogeny. Nat Commun, 2022, 13(1): 6968. doi: 10.1038/s41467-022-34630-w pmid: 36379955
[24]	Scheibenreif L, Littmann M, Orengo C, Rost B. FunFam protein families improve residue level molecular function prediction. BMC Bioinformatics, 2019, 20(1): 400. doi: 10.1186/s12859-019-2988-x pmid: 31319797
[25]	Ryu JY, Kim HU, Lee SY. Deep learning enables high-quality and high-throughput prediction of enzyme commission numbers. Proc Natl Acad Sci USA, 2019, 116(28): 13996-14001. doi: 10.1073/pnas.1821905116 pmid: 31221760
[26]	Han SR, Park M, Kosaraju S, Lee J, Lee H, Lee JH, Oh TJ, Kang M. Evidential deep learning for trustworthy prediction of enzyme commission number. Brief Bioinform, 2023, 25(1): bbad401.
[27]	Sanderson T, Bileschi ML, Belanger D, Colwell LJ. ProteInfer, deep neural networks for protein functional inference. eLife, 2023, 12: e80942.
[28]	Yu F, Koltun V. Multi-scale context aggregation by dilated convolutions. arXiv, 2015, doi:10.48550/arXiv.1511. 07122.
[29]	Kim GB, Kim JY, Lee JA, Norsigian CJ, Palsson BO, Lee SY. Functional annotation of enzyme-encoding genes using deep learning with transformer layers. Nat Commun, 2023, 14(1): 7370.
[30]	Vaswani A, Shazeer N, Parmar N, Uszkoreit J, Jones L, Gonez AN, Kaiser Ł, Polosukhin I. Attention is all you need. arXiv, 2017, doi:10.48550/arXiv.1706.03762.
[31]	Shi ZK, Deng R, Yuan QQ, Mao ZT, Wang RY, Li HR, Liao XP, Ma HW. Enzyme commission number prediction and benchmarking with hierarchical dual-core multitask learning framework. Research (Wash DC), 2023, 6: 0153.
[32]	Yu TH, Cui HY, Li JC, Luo YN, Jiang GD, Zhao HM. Enzyme function prediction using contrastive learning. Science, 379(6639): 1358-1363.
[33]	Huang W, Liu GQ. Study on hierarchical multi-label text classification method of MSML-BERT model. Computer Engineering and Applications, 2022, 58(15): 191-201. doi: 10.3778/j.issn.1002-8331.2111-0176
	黄伟, 刘贵全. MSML-BERT模型的层级多标签文本分类方法研究. 计算机工程与应用, 2022, 58(15): 191-201. doi: 10.3778/j.issn.1002-8331.2111-0176
[34]	Johnson R, Zhang T. Deep pyramid convolutional neural networks for text categorization. Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1:Long Papers), 2017, 562-570.
[35]	Tipton K. Translocases (EC 7): a new EC class. Enzyme Nomenclature News, 2018.

编辑推荐

Metrics

www.chinagene.cn
备案号：京ICP备09063187号-4
总访问:,今日访问:,当前在线:

数据集名称	蛋白质样本数量	EC编号数量
训练集	174287	2234
验证集	20753	1748
测试集	21742	1889
总计	216782	2234

基线网络	ResBlock	HFE	DP	Macro-F1	Micro-F1
Baseline				89.52±0.07	97.52±0.03
Baseline	√			92.76±0.18	98.42±0.04
Baseline	√	√		94.07±0.03	98.60±0.06
Baseline	√		√	93.84±0.18	98.62±0.03
Baseline	√	√	√	95.36±0.15	99.02±0.02

方法	Macro- Precision	Macro- Recall	Macro- F1	Micro- Precision	Micro- Recall	Micro- F1
ECPred	30.5	30.1	30.1	97.7	72.0	82.9
DeepEC	76.8	71.1	72.9	99.2	85.0	91.5
ECPICK	83.0	77.6	79.1	98.3	87.6	92.7
ProteInfer	87.7	85.4	86.1	99.5	93.3	96.3
DeepECtransformer	75.8	73.8	73.8	97.9	92.7	95.2
HDMLF	74.6	72.5	72.5	97.6	85.2	91.0
CLEAN	92.2	94.0	92.2	97.4	96.0	96.7
ECPN-HFGF	96.0	95.9	95.5	99.2	98.8	99.0

基于层级和全局特征结合的蛋白质序列EC编号预测

EC number prediction of protein sequences based on combination of hierarchical and global features

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 8

参考文献 35

相关文章 1

编辑推荐

Metrics