CoChat
Sign Up
Cited by 134Bioinformatics and Genomic NetworksGene expression and cancer classificationAI in cancer detection

Integrated multi-omics analysis of ovarian cancer using variational autoencoders

Open access
Hira, Muta Tah, Razzaque, M. A., Angione, Claudio, Scrivens, James, Sawan, Saladin, Sarker, Mosharraf·2021·Scientific Reports
DOI 10.1038/s41598-021-85285-4PMC PMC7973750
Open this paper in CoChat — free accountOpen PDF

Abstract

Cancer is a complex disease that deregulates cellular functions at various molecular levels (e.g., DNA, RNA, and proteins). Integrated multi-omics analysis of data from these levels is necessary to understand the aberrant cellular functions accountable for cancer and its development. In recent years, Deep Learning (DL) approaches have become a useful tool in integrated multi-omics analysis of cancer data. However, high dimensional multi-omics data are generally imbalanced with too many molecular features and relatively few patient samples. This imbalance makes a DL based integrated multi-omics analysis difficult. DL-based dimensionality reduction technique, including variational autoencoder (VAE), is a potential solution to balance high dimensional multi-omics data. However, there are few VAE-based integrated multi-omics analyses, and they are limited to pancancer. In this work, we did an integrated multi-omics analysis of ovarian cancer using the compressed features learned through VAE and an improved version of VAE, namely Maximum Mean Discrepancy VAE (MMD-VAE). First, we designed and developed a DL architecture for VAE and MMD-VAE. Then we used the architecture for mono-omics, integrated di-omics and tri-omics data analysis of ovarian cancer through cancer samples identification, molecular subtypes clustering and classification, and survival analysis. The results show that MMD-VAE and VAE-based compressed features can respectively classify the transcriptional subtypes of the TCGA datasets with an accuracy in the range of 93.2-95.5% and 87.1-95.7%. Also, survival analysis results show that VAE and MMD-VAE based compressed representation of omics data can be used in cancer prognosis. Based on the results, we can conclude that (i) VAE and MMD-VAE outperform existing dimensionality reduction techniques, (ii) integrated multi-omics analyses perform better or similar compared to their mono-omics counterparts, and (iii) MMD-VAE performs better than VAE in most omics dataset.

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1038/s41598-021-85285-4

Cites

Statistics

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • Maximum Mean Discrepancy as a regularization term outperforms Kullback-Leibler divergence.

    “Hira et al. [56] also found joint multimodal VAEs useful for fusing multi-omics data and support the findings of [41] that Maximum Mean Discrepancy as a regularization term outperforms the Kullback–Leibler divergence.”

    From Multimodal deep learning for biomedical data fusion: a review (2022) · cites this paper as [56]

    The Results (Dimensionality reduction; page unavailable) say MMD-VAE performed better than standard VAE on a majority of the ovarian-cancer datasets tested, but only suggest that the MMD-based loss function caused the difference. The Methods (VAE/MMD-VAE architecture; page unavailable) identify MMD and KL as their respective regularization terms. The unqualified claim that MMD regularization outperforms KL goes beyond those findings. (LLM, conf=0.87)

    Experimental — accuracy not yet measured

  • Hira et al. found joint multimodal VAEs useful for fusing multi-omics data.

    “Hira et al. [56] also found joint multimodal VAEs useful for fusing multi-omics data and support the findings of [41] that Maximum Mean Discrepancy as a regularization term outperforms the Kullback–Leibler divergence.”

    From Multimodal deep learning for biomedical data fusion: a review (2022) · cites this paper as [56]

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

The citation lists haven’t been fetched yet

References 61 · Cited by 134

Research works printed but not linked: 49

  1. [2]Torre, L. A. et al. Ovarian cancer statistics, 2018. CA68, 284–296 (2018). PMC6621554 29809280
  2. [3]Doubeni CA Doubeni AR Myers AE Diagnosis and management of ovarian cancer Am. Fam. Physician 2016 93 937 944 27281838
  3. [4]Rosenthal AN Menon U Jacobs IJ Screening for ovarian cancer Clin. Ostet. Gynecol. 2006 49 433 447 10.1097/00003081-200609000-00004 16885651 · DOI
  4. [5]Lu M Zhan X The crucial role of multiomic approach in cancer research and clinically relevant outcomes EPMA J. 2018 9 77 102 10.1007/s13167-018-0128-8 29515689 PMC5833337 · DOI
  5. [6]Hu R Wang X Zhan X Multi-parameter systematic strategies for predictive, preventive and personalised medicine in cancer EPMA J. 2013 4 2 10.1186/1878-5085-4-2 23339750 PMC3564825 · DOI
  6. [7]Cheng T Zhan X Pattern recognition for predictive, preventive, and personalized medicine in cancer EPMA J. 2017 8 51 60 10.1007/s13167-017-0083-9 28620443 PMC5471804 · DOI
  7. [8]Zhan, X., Long, Y. & Lu, M. Exploration of variations in proteome and metabolome for predictive diagnostics and personalized treatment algorithms: Innovative approach and examples for potential clinical application. J. Proteom.188, 30 – 40, 10.1016/j.jprot.2017.08.020 (2018). The key-role of multi-omics in predictive, preventive and personalised medicine as the medicine of the future. 28851587 · DOI
  8. [10]Chakraborty S Hosen MI Ahmed M Shekhar HU Onco-multi-omics approach: A new frontier in cancer research BioMed Res. Int. 2018 2018 9836256 9836256 10.1155/2018/9836256 30402498 PMC6192166 · DOI
  9. [11]Olivier M Asmis R Hawkins GA Howard TD Cox LA The need for multi-omics biomarker signatures in precision medicine Int. J. Mol. Sci. 2019 20 4781 10.3390/ijms20194781 PMC6801754 31561483 · DOI
  10. [12]Wu C A selective review of multi-level omics data integration using variable selection High-Throughput 2019 8 4 10.3390/ht8010004 PMC6473252 30669303 · DOI
  11. [13]Bell D Integrated genomic analyses of ovarian carcinoma Nature 2011 474 609 615 10.1038/nature10166 21720365 PMC3163504 · DOI · Paper page
  12. [15]Zhang H Integrated proteogenomic characterization of human high-grade serous ovarian cancer Cell 2016 166 755 765 10.1016/j.cell.2016.05.069 27372738 PMC4967013 · DOI · Paper page
  13. [17]Zhang, Q. et al. Integrated multiomic analysis reveals comprehensive tumour heterogeneity and novel immunophenotypic classification in hepatocellular carcinomas. Gut68, 2019–2031, 10.1136/gutjnl-2019-318912 (2019). https://gut.bmj.com/content/68/11/2019.full.pdf. PMC6839802 31227589 · DOI
  14. [18]Liu Q Hu P Association analysis of deep genomic features extracted by denoising autoencoders in breast cancer Cancers 2019 11 494 10.3390/cancers11040494 PMC6520782 30959966 · DOI
  15. [19]Weinstein JN The cancer genome atlas pan-cancer analysis project Nat. Genet. 2013 45 1113 10.1038/ng.2764 24071849 PMC3919969 · DOI · Paper page
  16. [20]Champion M Module analysis captures pancancer genetically and epigenetically deregulated cancer driver genes for smoking and antiviral response EBioMedicine 2018 27 156 166 10.1016/j.ebiom.2017.11.028 29331675 PMC5828545 · DOI
  17. [22]Gonzalez-Reymundez, A. & Vazquez, A. I. Multi-omic signatures identify pan-cancer classes of tumors beyond tissue of origin. bioRxiv10.1101/806323 (2019). PMC7239905 32433524
  18. [23]Singh A DIABLO: An integrative approach for identifying key molecular drivers from multi-omics assays Bioinformatics 2019 35 3055 3062 10.1093/bioinformatics/bty1054 30657866 PMC6735831 · DOI
  19. [24]Terunuma A Myc-driven accumulation of 2-hydroxyglutarate is associated with breast cancer prognosis J. Clin. Investig. 2014 124 398 412 10.1172/JCI71180 24316975 PMC3871244 · DOI · Paper page
  20. [25]Zhang Z Molecular subtyping of serous ovarian cancer based on multi-omics data Sci. Rep. 2016 6 26001 10.1038/srep26001 27184229 PMC4868982 · DOI
  21. [26]Fu A Chang HR Zhang Z-F Integrated multiomic predictors for ovarian cancer survival Carcinogenesis 2018 39 860 868 10.1093/carcin/bgy055 29897425 PMC6031025 · DOI
  22. [27]Papp E Integrated genomic, epigenomic, and expression analyses of ovarian cancer cell lines Cell Rep. 2018 25 2617 2633 10.1016/j.celrep.2018.10.096 30485824 PMC6481945 · DOI
  23. [28]Wang, Z. & Wang, Y. Exploring dna methylation data of lung cancer samples with variational autoencoders. In 2018 IEEE International Conference on Bioinformatics and Biomedicine (BIBM) 1286–1289, 10.1109/BIBM.2018.8621365 (2018). · DOI
  24. [29]Zhang, X. et al. Integrated multi-omics analysis using variational autoencoders: Application to pan-cancer classification. arXiv preprint arXiv:1908.06278 (2019). · arXiv
  25. [30]Fakoor, R., Ladhak, F., Nazi, A. & Huber, M. Using deep learning to enhance cancer diagnosis and classification. In Proceedings of the international conference on machine learning, vol. 28 (ACM New York, USA, 2013).
  26. [31]Kingma, D. P. & Welling, M. Auto-encoding variational bayes. arXiv preprint arXiv:1312.6114 (2013). · arXiv
  27. [32]Titus, A. J., Wilkins, O. M., Bobak, C. A. & Christensen, B. C. An unsupervised deep learning framework with variational autoencoders for genome-wide dna methylation analysis and biologic feature extraction applied to breast cancer. bioRxiv 433763 (2018).
  28. [33]George, T. M. & Lio, P. Unsupervised machine learning for data encoding appliedto ovarian cancer transcriptomes. BioRxiv855593 (2019).
  29. [34]Way, G. P. & Greene, C. S. Evaluating deep variational autoencoders trained on pan-cancer gene expression. arXiv preprint arXiv:1711.04828 (2017). · arXiv
  30. [35]Palazzo M Beauseroy P Yankilevich P A pan-cancer somatic mutation embedding using autoencoders BMC Bioinform. 2019 20 655 10.1186/s12859-019-3298-z PMC6907172 31829157 · DOI
  31. [36]Zhao, S., Song, J. & Ermon, S. Infovae: Information maximizing variational autoencoders. arXiv preprint arXiv:1706.02262 (2017). · arXiv
  32. [37]Basu, T. & Murthy, C. Effective text classification by a supervised feature selection approach. In 2012 ieee 12th international conference on data mining workshops 918–925 (IEEE, 2012).
  33. [38]Chao G Luo Y Ding W Recent advances in supervised dimension reduction: A survey Mach. Learn. Knowl. Extract. 2019 1 341 358 10.3390/make1010020 · DOI
  34. [45]Han J Pei J Kamber M Data Mining: Concepts and Techniques 2011 Amsterdam Elsevier
  35. [47]Kullback S Leibler RA On information and sufficiency Ann. Math. Stat. 1951 22 79 86 10.1214/aoms/1177729694 · DOI · Paper page
  36. [48]Chen, X. et al. Variational lossy autoencoder. arXiv preprint arXiv:1611.02731 (2016). · arXiv
  37. [49]Sønderby, C. K., Raiko, T., Maaløe, L., Sønderby, S. K. & Winther, O. Ladder variational autoencoders. Adv. Neural Inf. Process. Syst.3738–3746 (2016).
  38. [50]Gretton, A., Borgwardt, K. M., Rasch, M., Schölkopf, B. & Smola, A. J. A kernel method for the two-sample-problem. In Proceedings of the 19th International Conference on Neural Information Processing Systems, NIPS’06, 513-520 (MIT Press, Cambridge, MA, USA, 2006).
  39. [52]Ioffe, S. & Szegedy, C. Batch normalization: Accelerating deep network training by reducing internal covariate shift (2015). arXiv:1502.03167. · arXiv
  40. [53]Tothill RW Novel molecular subtypes of serous and endometrioid ovarian cancer linked to clinical outcome Clin. Cancer Res. 2008 14 5198 5208 10.1158/1078-0432.CCR-08-0196 18698038 · DOI
  41. [54]Braicu EI Dynamic of molecular subtypes of high-grade serous ovarian cancer in paired primary and relapsed biopsies J. Clin. Oncol. 2019 37 e17091 e17091 10.1200/JCO.2019.37.15_suppl.e17091 · DOI
  42. [55]Shen H Laird PW Interplay between the cancer genome and epigenome Cell 2013 153 38 55 10.1016/j.cell.2013.03.008 23540689 PMC3648790 · DOI
  43. [56]Sun W The association between copy number aberration, DNA methylation and gene expression in tumor samples Nucleic Acids Res. 2018 46 3009 3018 10.1093/nar/gky131 29529299 PMC5887505 · DOI
  44. [57]Damgacioglu H Celik E Celik N Estimating gene expression from high-dimensional DNA methylation levels in cancer data: A bimodal unsupervised dimension reduction algorithm Comput. Ind. Eng. 2019 130 348 357 10.1016/j.cie.2019.02.038 · DOI
  45. [62]Harrell FE Jr Lee KL Mark DB Multivariable prognostic models: Issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors Stat. Med. 1996 15 361 387 10.1002/(SICI)1097-0258(19960229)15:4<361::AID-SIM168>3.0.CO;2-4 8668867 · DOI
  46. [63]Raykar, V. C., Steck, H., Krishnapuram, B., Dehing-Oberije, C. & Lambin, P. On ranking in survival analysis: Bounds on the concordance index. In Proceedings of the 20th International Conference on Neural Information Processing Systems, NIPS’07, 1209-1216 (Curran Associates Inc., Red Hook, NY, USA, 2007).
  47. [66]Gerds TA Schumacher M Consistent estimation of the expected brier score in general survival models with right-censored event times Biom. J. 2006 48 1029 1040 10.1002/bimj.200610301 17240660 · DOI
  48. [68]Kingma, D. P. & Ba, J. Adam: A Method for Stochastic Optimization (2014). arXiv:1412.6980. · arXiv
  49. [69]Kang F Ndrg2 gene expression pattern in ovarian cancer and its specific roles in inhibiting cancer cell proliferation and suppressing cancer cell apoptosis J. Ovarian Res. 2020 13 1 10 10.1186/s13048-020-00649-0 PMC7189606 32345304 · DOI

Authors
Hira, M. T., Razzaque, M. A., Angione, C., Scrivens, J., Sawan, S., & Sarker, M.
Affiliations
  • Teesside University (GB)
  • James Cook University Hospital (GB)
Corresponding author
Razzaque, M. A.
ORCID
Venue
Scientific Reports
Publisher
Springer Science and Business Media LLC
Published
2021-03-18
Volume
11
Issue
1
Pages
6265-6265
ISSN
2045-2322 (ISSN-L)
DOI
10.1038/s41598-021-85285-4
PMC
PMC7973750
OpenAlex
W3138022488
Cited by
134
References
61
Citations by year
2026: 26 · 2025: 27 · 2024: 31 · 2023: 26 · 2022: 16 · 2021: 8
FWCI
6.44
Type
journal article
Language
en
Update notices
correction (2021-08-11)
Open access
Licence
Locations
  • Scientific Reports (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • TeesRep (Teesside University) (publishedVersion, cc-by)
  • Liverpool John Moores University (acceptedVersion)
  • Teesside University Research Portal (Teesside University) (submittedVersion, cc-by)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
Teesside University
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Bioinformatics and Genomic Networks
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Gene expression and cancer classification
  • Physical Sciences › Computer Science › Artificial Intelligence › AI in cancer detection
Keywords
ovarian cancer, multi-omics integration, variational autoencoder, dimensionality reduction, tumor subtypes, survival analysis, cancer prognosis, TCGA
MeSH terms
Deep Learning, Data Analysis, Female, Humans, Ovarian Neoplasms, Prognosis, Cohort Studies, Cluster Analysis, Computational Biology, Gene Expression Profiling, Genomics, Epigenesis, Genetic, Kaplan-Meier Estimate, Epigenomics, Transcriptome

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review