癌症多组学数据深度自编码器整合分型方法

doi:10.3778/j.issn.1002-8331.2102-0146

摘要/Abstract

摘要： 在癌症研究中，随着高通量测序技术发展已经产生了海量的复杂数据。尽管有了一些利用深度学习和统计学方法进行多组学数据整合的研究，但目前仍缺乏较为有效率的整合方法。因此提出一种基于深度自编码器的多组学数据整合方法（deep autoencoder for multi-omics integration，DAEMI）。它利用自编码器中的瓶颈层，学习多组学数据的特征表示。与先前利用深度学习整合的研究相比，DAEMI可以发现明显生存差异的癌症亚型。同时因为不需要生存数据来选择特征，DAEMI可以使用更多特征进行[K]均值聚类，进而完成癌症分型任务。将DAEMI应用于模拟数据集与四个癌症数据集实验，通过与高阶路径相似度网络的融合模型（HOPES）、相似性网络融合（SNF）、iClusterPlus和moCluster进行比较，结合模拟数据集测试结果与真实癌症数据集测试结果来看，DAEMI要优于其他方法。相应的生物功能分析揭示，神经退行性疾病与线粒体功能障碍可能与癌症共享某些生物学通路。

关键词: 多组学数据整合, 癌症分型, [K]均值, 深度学习, 生存分析

Abstract: In cancer research, high-throughput sequencing techniques have yielded a large amount of complex heterogeneous data. Although several deep learning or statistical methods have been applied to integrate such data, there is a lack of work on how to integrate multi-omics data more efficiently. Therefore, a deep learning-based multi-omics data integration method named deep autoencoder for multi-omics integration（DAEMI） is proposed. It utilizes a deep learning algorithm named autoencoder, which uses a bottleneck layer in the network to get a compressed knowledge representation of the original input. Unlike previous deep learning integration studies, this method finds more subtypes with significant survival differences. DAEMI does not rely on survival data for compressed features selection. It uses all features obtained from the bottleneck layer and uses [K]-means to identify cancer subtypes. By comparing with high-order path elucidated similarity（HOPES）, Similarity network fusion（SNF）, iClusterPlus, and moCluster on 4 cancer datasets and simulate dataset, DAEMI performs better than other methods. Functional analysis revealed neurodegenerative diseases and mitochondrial dysfunction may share some pathways with cancer.

Key words: multi-omics data integration, cancer subtyping, [K]-means, deep learning, survival analysis

曹业伟, 刘飞. 癌症多组学数据深度自编码器整合分型方法[J]. 计算机工程与应用, 2022, 58(18): 154-161.

CAO Yewei, LIU Fei. Multi-Omics Data Deep Autoencoder Integration for Cancer Subtyping[J]. Computer Engineering and Applications, 2022, 58(18): 154-161.

参考文献

[1] KRISTENSEN V N，LINGJ?RDE O C，RUSSNES H G，et al.Principles and methods of integrative genomic analyses in cancer[J].Nature Reviews Cancer，2014，14（5）：299-313.
[2] HUANG S，CHAUDHARY K，GARMIRE L X.More is better：recent progress in multi-omics data integration methods[J].Frontiers in Genetics，2017，8.
[3] XU A，CHEN J，PENG H，et al.Simultaneous interrogation of cancer omics to identify subtypes with significant clinical differences[J].Frontiers in Genetics，2019，10.
[4] FORBES S A，BINDAL N，BAMFORD S，et al.COSMIC：mining complete cancer genomes in the catalogue of somatic mutations in cancer[J].Nucleic Acids Research，2010，39：945-950.
[5] CONSORTIUM I C G.International network of cancer genome projects[J].Nature，2010，464（7291）：993.
[6] WEINSTEIN J N，COLLISSON E A，MILLS G B，et al.The cancer genome atlas pan-cancer analysis project[J].Nature Genetics，2013，45（10）：1113-1120.
[7] PUGH T J，MOROZOVA O，ATTIYEH E F，et al.The genetic landscape of high-risk neuroblastoma[J].Nature Genetics，2013，45（3）：279-284.
[8] MO Q，WANG S，SESHAN V E，et al.Pattern discovery and cancer gene identification in integrated cancer genomic data[J].Proceedings of the National Academy of Sciences，2013，110（11）：4245-4250.
[9] WANG B，MEZLINI A M，DEMIR F，et al.Similarity network fusion for aggregating data types on a genomic scale[J].Nature Methods，2014，11（3）：333-337.
[10] KIM D，LI R，LUCAS A，et al.Using knowledge-driven genomic interactions for multi-omics data analysis：metadimensional models for predicting clinical outcomes in ovarian carcinoma[J].Journal of the American Medical Informatics Association，2016：165.
[11] CHAUDHARY K，POIRION O B，LU L，et al.Deep learning-based multi-omics integration robustly predicts survival in liver cancer[J].Clinical Cancer Research，2017，24（6）：1248-1259.
[12] ZHANG L，LV C，JIN Y，et al.Deep learning-based multi-omics data integration reveals two prognostic subtypes in high-risk neuroblastoma[J].Frontiers in Genetics，2018，9.
[13] ZHAO Z，LI Y，WU Y，et al.Deep learning-based model for predicting progression in patients with head and neck squamous cell carcinoma[J].Cancer Biomarkers，2019，27（1）：19-28.
[14] CHAI H，ZHOU X，CUI Z，et al.Integrating multi-omics data with deep learning for predicting cancer prognosis[J].Cold Spring Harbor Laboratory，2019.
[15] JONATHAN R，SIKANDER H，ALTUNA A.Evaluation of colorectal cancer subtypes and cell lines using deep learning[J].Life Science Alliance，2019，2（6）.
[16] MENG C，HELM D，FREJNO M，et al.MoCluster：identifying joint patterns across multiple omics data sets[J].Journal of Proteome Research，2015，15（3）：755-765.
[17] WAY G P，GREENE C S.Extracting a biologically relevant latent space from cancer transcriptomes with variational autoencoders[J].BioRxiv，2017：174474.
[18] BENGIO Y.Learning deep architectures for AI[J].Foundations and Trends? in Machine Learning，2009，2（1）：1-127.
[19] CHOLLET F.Keras[M/OL]//GitHub Repository.GitHub，2015.https：//github.com/fchollet/keras.
[20] KINGMA D P，BA J.Adam：a method for stochastic optimization[J].arXiv：1412.6980，2014.
[21] STEINLEY D.K-means clustering：a half-century synthesis[J].British Journal of Mathematical and Statistical Psychology，2006，59（1）：1-34.
[22] PEDREGOSA F，VAROQUAUX G，GRAMFORT A，et al.Scikit-learn：machine learning in python[J].The Journal of Machine Learning Research，2011，12：2825-2830.
[23] KAPLAN E L，MEIER P.Nonparametric estimation from incomplete observations[J].Journal of the American Statistical Association，1958，53（282）：457-481.
[24] MANTEL N.Evaluation of survival data and two new rank order statistics arising in its consideration[J].Cancer Chemotherapy Reports，1966，50（3）：1623-1670.
[25] DAVIDSON-PILON C，KALDERSTAM J，JACOBSON N，et al.CamDavidsonPilon/Lifelines：V0.24.13[R/OL].Zenodo，2020.https：//doi.org/10.5281/zenodo.3903636.DOI：10.5281/
zenodo.3903636.
[26] LENG N，DAWSON J，KENDZIORSKI C.EBSeq：An R package for differential expression analysis using RNA-Seq data[J].R package Version，2015，1（10）.
[27] ELLINGER J，GROMES A，POSS M，et al.Systematic expression analysis of the mitochondrial complex III subunits identifies UQCRC1 as biomarker in clear cell renal cell carcinoma[J].Oncotarget，2016，7（52）：86490-86499.
[28] THOMPSON J A，MARSIT C J.A methylation-to-expression feature model for generating accurate prognostic risk scores and identifying disease targets in clear cell kidney cancer[C]//Pacific Symposium on Biocomputing，2017：509-520.
[29] GOLDENRING J，SHEN K R，VAUGHAN H D，et al.Identification of a small GTP-binding protein，Rab25，expressed in the gastrointestinal mucosa，kidney，and lung[J].Journal of Biological Chemistry，1993，268（25）：18419-18422.
[30] WANG W，XIE G，REN Z，et al.Gene selection for the discrimination of colorectal cancer[J].Current Molecular Medicine，2019，19.
[31] YU G，WANG L G，HAN Y，et al.Cluster profiler：an R package for comparing biological themes among gene clusters[J].OMICS：A Journal of Integrative Biology，2012，16（5）：284-287.
[32] KANEHISA M，GOTO S，SATO Y，et al.KEGG for integration and interpretation of large-scale molecular data sets[J].Nucleic Acids Research，2011，40（1）：109-114.
[33] ASHBURNER M，BALL C A，BLAKE J A，et al.Gene ontology：tool for the unification of biology[J].Nature Genetics，2000，25（1）：25-29.
[34] ZONG W X，RABINOWITZ J D，WHITE E.Mitochondria and cancer[J].Molecular Cell，2016，61（5）：667-676.
[35] SOLAINI G，SGARBI G，BARACCA A.Oxidative phosphorylation in cancer cells[J].Biochimica et Biophysica Acta（BBA）-Bioenergetics，2011，1807（6）：534-542.
[36] de MOURA M B，DOS SANTOS L S，HOUTEN B V.Mitochondrial dysfunction in neurodegenerative diseases and cancer[J].Environmental and Molecular Mutagenesis，2010.
[37] THANAN R，OIKAWA S，HIRAKU Y，et al.Oxidative stress and its significant roles in neurodegenerative diseases and cancer[J].International Journal of Molecular Sciences，2014，16（1）：193-217.