整合磷酸蛋白質組數據與深度學習的激酶活性圖譜預測與研究

以作者查詢圖書館館藏

、以作者查詢臺灣博碩士

、以作者查詢全國書目

、勘誤回報

、線上人數：105

、訪客IP：3.145.164.139

姓名

何名豐(Ming-Feng Ho) 查詢紙本館藏

畢業系所

資訊工程學系

論文名稱

整合磷酸蛋白質組數據與深度學習的激酶活性圖譜預測與研究
(Integration of Phosphoproteomics Data and Deep Learning for Predicting and Investigating Kinase Activity Profiles)

相關論文

★ 空氣汙染物與疾病關聯性之研究與利用深度學習預測疾病	★ 利用質譜儀資料快速檢測金黃色葡萄球菌之抗藥性
★ 根據質譜儀資料辨識大腸桿菌抗藥性之特徵峰值	★ 蛋白質賴氨酸丙二酰化修飾作用位點之預測系統
★ 基於機器學習方法的抗微生物肽活性預測及特徵分析	★ 用於預測抗菌肽多種功能類別的多標籤分類器
★ 利用機器學習預測濁水溪沖積扇區域之地下水砷汙染	★ 基於質譜儀資料使用深度學習方法預測不同地區之耐甲氧西林金黃色葡萄球菌之抗藥性
★ EnHemo：融合蛋白質語言模型的集成框架用於識別高活性抗菌肽的溶血毒性

檔案

[Endnote RIS 格式]

[Bibtex 格式]

[相關文章]

[文章引用]

[完整記錄]

[館藏目錄]

至系統瀏覽論文 ( 永不開放)

摘要(中)

激酶是一類生物分子，通常是蛋白質，其主要功能是在細胞內催化特定的化學反應。這些反應對於代謝、信號傳導和細胞增長等關鍵細胞過程至關重要。激酶活性指的是這些催化反應的速率，這種活性的變化可以直接影響細胞功能和狀態。因此，理解激酶活性對於闡明疾病機制和開發治療策略至關重要。激酶活性可以通過基於質譜的磷酸蛋白質組學數據進行識別。目前，已有多種工具可用於預測激酶活性，但這些工具在數據要求方面存在顯著差異，其中一些工具需要複雜且難以獲得的數據。本研究旨在設計一個模型，使用磷酸化蛋白數據來預測激酶表達水平，從而預測激酶活性圖譜。我們的方法採用深度學習方法作為框架建立預測模型，並使用不同架構的模型，包含深度神經網路(Deep Neural Networks，DNN)以及卷積神經網路（Convolutional Neural Networks，CNN），還有不同特徵資料，例如激酶底物資料和磷酸化蛋白資料，來作為模型輸入，通過比較來觀察哪種模型架構以及特徵種類的組合更適合激酶活性譜的預測。對於深度學習模型，我們對不同模型層數、神經元數量以及特徵種類進行測試。實驗結果表明，使用三層卷積層，神經元數量為(32,16,8)的CNN模型，搭配經過特徵選擇方法的磷酸化蛋白資料作為特徵，這個模型架構組合，於所有測試的模型架構中取得最好的結果。在與其他方法比較的部分，在斯皮爾曼排名相關係數的部分，取得0.4101，高於其他方法的0.0655和0.0080；在C-index的部分，在閥值x大於9之後取得優於其他方法的結果，整體來說也相對穩定。

摘要(英)

Kinases are biomolecules, typically proteins, whose primary function is to catalyze specific chemical reactions within cells, essential for processes such as metabolism, signal transduction, and cell growth. Kinase activity, referring to the rate of these reactions, can directly affect cellular function and state, making it crucial for understanding disease mechanisms and developing therapeutic strategies. Kinase activity can be identified using mass spectrometry-based phosphoproteomics data. Various tools exist for predicting kinase activity, but they differ significantly in data requirements, with some necessitating complex and hard-to-obtain data. This study aims to design a model that predicts kinase expression levels using phosphorylation protein data to forecast kinase activity profiles. We employed deep learning methods, constructing predictive models using different architectures, including Deep Neural Networks (DNN) and Convolutional Neural Networks (CNN), and various feature types, such as kinase-substrate data and phosphorylation protein data. By comparing these combinations, we sought to identify the most suitable model architecture and feature type for predicting kinase activity profiles. Our experimental results indicate that a CNN model with three convolutional layers and neuron counts of 32, 16, and 8, using phosphorylation protein data refined through feature selection methods, achieved the best results among all tested model architectures. In the comparative analysis with other methods, our model achieved a Spearman rank correlation coefficient of 0.4101, surpassing the other methods, which scored 0.0655 and 0.0080. Regarding the C-index, our model outperformed the other methods when the threshold x exceeded 9 and demonstrated overall relative stability.

關鍵字(中)

★ 磷酸蛋白質組數據
★ 激酶活性圖譜
★ 深度學習

關鍵字(英)

★ Phosphoproteomics data
★ Kinase activity profiles
★ Deep learning

論文目次

中文摘要. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I
Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . II
致謝. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . III
TableofContents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IV
ListofFigures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VI
ListofTables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VII
1 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Background. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 RelatedWorks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3MotivationandGoal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2 MaterialsandMethods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.1 PhosphorylationDataCollection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.2 Kinase-substrateDataCollection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.3 DataPreprocessing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.4ModelArchitectures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.4.1 DeepNeuralNetworks(DNN) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.4.2 ConvolutionNeuralNetworks(CNN) . . . . . . . . . . . . . . . . . . . . . . . . 9
2.5 FeatureSelection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.6 EvaluationMetrics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.6.1MSE,MAEandRsquared . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.6.2 Rank-basedCorrelationCoefficient . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.6.3 C-index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.7 FeatureImportance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.8 CorrelationMatrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
3.1 PerformanceComparisonofDifferentModelArchitectures:DNNvs.CNN. . . . 14
3.2 PerformanceComparisonUsingDifferentFeatures:Kinase-substrateDatavs.All
SitesConsideredasFeatureswithorwithoutFeatureSelection . . . . . . . . . . . . 18
3.3 ComparisonwithOtherMethods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
4 DiscussionsandConclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
4.1 Discussions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
4.2 Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

參考文獻

[1] P. Cohen, “The role of protein phosphorylation in neural and hormonal control of cellular
activity,” Nature, vol. 296, no. 5858, pp. 613–620, 1982.
[2] G.Manning,D.B.Whyte,R.Martinez,T.Hunter,andS.Sudarsanam,“Theproteinkinase
complement of the human genome,” Science, vol. 298, no. 5600, pp. 1912–1934, 2002.
[3] D. D. Wiredja, M. Koyutürk, and M. R. Chance, “The ksea app: A web-based tool for
kinase activity inference from quantitative phosphoproteomics,” Bioinformatics, vol. 33,
no. 21, pp. 3489–3491, 2017.
[4] R. Beekhof et al., “Inka, an integrative data analysis pipeline for phosphoproteomic in
ference of active kinases,” Molecular Systems Biology, vol. 15, no. 4, e8250, 2019.
[5] P. Casado and P. R. Cutillas, “A self-validating quantitative mass spectrometry method
for assessing the accuracy of high-content phosphoproteomic experiments,” Molecular &
Cellular Proteomics, vol. 10, no. 1, 2011.
[6] H.-D.Huang, T.-Y. Lee, S.-W. Tzeng, and J.-T. Horng, “Kinasephos: A web tool for iden
tifying protein kinase-specific phosphorylation sites,” Nucleic Acids Research, vol. 33,
no. suppl_2, W226–W229, 2005.
[7] S.Yılmaz, M. Ayati, D. Schlatzer, A. E. Çiçek, M. R. Chance, and M. Koyutürk, “Robust
inference of kinase activity using functional networks,” Nature Communications, vol. 12,
no. 1, p. 1177, 2021.
[8] P. Casado et al., “Kinase-substrate enrichment analysis provides insights into the het
erogeneity of signaling pathway activation in leukemia cells,” Science Signaling, vol. 6,
no. 268, rs6–rs6, 2013.
[9] S.R.Piersma,A.Valles-Marti, F. Rolfs, T. V. Pham, A.A.Henneman,andC.R.Jiménez,
“Inferring kinase activity from phosphoproteomic data: Tool comparison and recent ap
plications,” Mass Spectrometry Reviews, e21808, 2022.
[10] E. H. Wilkes, P. Casado, V. Rajeeve, and P. R. Cutillas, “Kinase activity ranking using
phosphoproteomics data (karp) quantifies the contribution of protein kinases to the regu
lation of cell viability,” Molecular & Cellular Proteomics, vol. 16, no. 9, pp. 1694–1704,
2017.
[11] M.V.Kuleshovetal., “Kea3: Improved kinase enrichment analysis via data integration,”
Nucleic Acids Research, vol. 49, no. W1, W304–W316, 2021.
[12] S. Crowl, B. T. Jordan, H. Ahmed, C. X. Ma, and K. M. Naegle, “Kstar: An algorithm to
predict patient-specific kinase activities from phosphoproteomic data,” Nature Commu
nications, vol. 13, no. 1, p. 4283, 2022.
[13] S.Vasaikaretal.,“Proteogenomicanalysisofhumancoloncancerrevealsnewtherapeutic
opportunities,” Cell, vol. 177, no. 4, pp. 1035–1049, 2019.
[14] T. Zhang et al., “Interrogating kinase–substrate relationships with proximity labeling and
phosphorylation enrichment,” Journal of Proteome Research, vol. 21, no. 2, pp. 494–506,
2022.
[15] M. Hijazi, R. Smith, V. Rajeeve, C. Bessant, and P. R. Cutillas, “Reconstructing kinase
network topologies from phosphoproteomics data reveals cancer-associated rewiring,”
Nature Biotechnology, vol. 38, no. 4, pp. 493–502, 2020.
[16] J. Adachi et al., “Systematic identification of alk substrates by integrated phosphopro
teome and interactome analysis,” Life Science Alliance, vol. 5, no. 8, 2022.
[17] N. Sugiyama, H. Imamura, and Y. Ishihama, “Large-scale discovery of substrates of the
human kinome,” Scientific Reports, vol. 9, no. 1, p. 10503, 2019.
[18] R. Ma, S. Li, W. Li, L. Yao, H.-D. Huang, and T.-Y. Lee, “Kinasephos 3.0: Redesign and
expansion of the prediction on kinase-specific phosphorylation sites,” Genomics, Pro
teomics and Bioinformatics, vol. 21, no. 1, pp. 228–241, 2023.
[19] R. Linding et al., “Networkin: A resource for exploring cellular phosphorylation net
works,” Nucleic Acids Research, vol. 36, no. suppl_1, pp. D695–D699, 2007.
[20] P. V. Hornbeck et al., “Phosphositeplus: A comprehensive resource for investigating the
structure and function of experimentally determined post-translational modifications in
man and mouse,” Nucleic Acids Research, vol. 40, no. D1, pp. D261–D270, 2012.
[21] P. Lo Surdo et al., “Signor 3.0, the signaling network open resource 3.0: 2022 update,”
Nucleic Acids Research, vol. 51, no. D1, pp. D631–D637, 2023.
[22] K.O’sheaandR.Nash,“Anintroductiontoconvolutionalneuralnetworks,”arXivpreprint
arXiv:1511.08458, 2015.
[23] V. Sze, Y.-H. Chen, T.-J. Yang, and J. S. Emer, “Efficient processing of deep neural net
works: A tutorial and survey,” Proceedings of the IEEE, vol. 105, no. 12, pp. 2295–2329,
2017.
[24] B. Nuche-Berenguer, I. Ramos-Álvarez, and R. Jensen, “The p21-activated kinase, pak2,
is important in the activation of numerous pancreatic acinar cell signaling cascades and in the onset of early pancreatitis events,” Biochimica et Biophysica Acta (BBA)-Molecular
Basis of Disease, vol. 1862, no. 6, pp. 1122–1136, 2016.
[25] M. Alfaidi, U. Bhattarai, and A. W. Orr, “Nck1, but not nck2, mediates disturbed flow
induced p21-activated kinase activation and endothelial permeability,” Journal of the
American Heart Association, vol. 9, no. 11, e016099, 2020.
[26] KEGG PATHWAY: Adrenergic signaling in cardiomyocytes- Homo sapiens (human) —
genome.jp, https://www.genome.jp/pathway/hsa04261, [Accessed 05-06-2024].
[27] Harmonizome — maayanlab.cloud, https://maayanlab.cloud/Harmonizome/, [Ac
cessed 05-06-2024].
[28] L.-I.Hsuetal.,“Pathwayanalysisofgenome-wideassociationstudyinchildhoodleukemia
amonghispanics,”CancerEpidemiology,Biomarkers&Prevention,vol.25,no.5,pp.815
822, 2016.
[29] X. Ni et al., “Long non-coding rna zeb1-as1 promotes colon adenocarcinoma malignant
progression via mir-455-3p/pak2 axis,” Cell Proliferation, vol. 53, no. 1, e12723, 2020.
[30] J. Luo et al., “Establishment of an immune-related gene pair model to predict colon ade
nocarcinoma prognosis,” BMC cancer, vol. 20, pp. 1–11, 2020.
[31] C.Hofmann,M.Shepelev,andJ.Chernoff,“Thegeneticsofpak,”Journalofcell science,
vol. 117, no. 19, pp. 4343–4354, 2004.
[32] E. Sementino et al., “Inactivation of p21-activated kinase 2 (pak2) inhibits the develop
ment of nf2-deficient tumors by restricting downstream hedgehog and wnt signaling,”
Molecular Cancer Research, vol. 20, no. 5, pp. 699–711, 2022.
[33] R. Santoro, C. Carbone, G. Piro, P. J. Chiao, and D. Melisi, “Tak-ing aim at chemore
sistance: The emerging role of map3k7 as a target for cancer therapy,” Drug Resistance
Updates, vol. 33, pp. 36–42, 2017.
[34] S. Ikram, A. Rege, M. Y. Negesse, A. G. Casanova, N. Reynoird, and E. M. Green, “The
smyd3-map3k2 signaling axis promotes tumor aggressiveness and metastasis in prostate
cancer,” Science Advances, vol. 9, no. 46, eadi5921, 2023.
[35] K. Nguyen et al., “Nek family review and correlations with patient survival outcomes in
various cancer types,” Cancers, vol. 15, no. 7, p. 2067, 2023.

指導教授

洪炯宗吳立青(Jorng-Tzong Horng Li-Ching Wu)

審核日期

2024-8-5

推文