Pith. sign in

REVIEW 4 major objections 6 minor 24 references

Integrated Multi-omics Analysis Using Variational Autoencoders: Application to Pan-cancer Classification

T0 review · 4 major / 6 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read OmiVAE, an end-to-end variational autoencoder plus classifier, reaches 97.49% average accuracy across 33 tumour types and normal samples from a pan-cancer multi-omics dataset, and beats single-omics models.

desk verdict OmiVAE is a genuinely useful practical multi-omics classifier, but the central accuracy numbers hinge on an unstated cross-validation nesting assumption that the paper never confirms. read the letter →

arxiv 1908.06278 v1 pith:VZZTB3GI submitted 2019-08-17 cs.LG q-bio.GNstat.ML

Xiaoyu Zhang , Jingqing Zhang , Kai Sun , Xian Yang , Chengliang Dai , Yike Guo This is my paper · ORCID
classification cs.LGq-bio.GNstat.ML
keywords variationalautoencoderpan-cancerclassificationmulti-omicsintegrationDNAmethylationgeneexpressiondeeplearninglatentspacegenomics
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

OmiVAE is built to solve a mismatch: cancer omics profiles have tens of thousands to hundreds of thousands of molecular features but only thousands of labelled samples, so classifiers overfit. The paper's proposal is to train a variational autoencoder first without labels, letting a low-dimensional latent space organise samples into a hierarchy that mirrors organs, subtypes, and normal tissue, then attach a classifier to that latent space and fine-tune with labels. On 9,081 pan-cancer samples covering 33 tumour types plus normal tissue, the end-to-end model reaches 97.49% average accuracy after stratified 10-fold cross-validation, and 97.88% when normal samples are excluded. In both its unsupervised and end-to-end modes, combining DNA methylation with gene expression outperforms using either alone. If the result holds, integrated multi-omics classification can be improved by a single architecture that does representation learning and classification together, rather than by manual feature selection followed by a separate classifier.

What carries the argument

The load-bearing machine is the joint loss $\mathcal{L}_{\mathrm{total}} = \alpha \mathcal{L}_{\mathrm{vae}} + \beta \mathcal{L}_{\mathrm{class}}$, coupling a variational-autoencoder reconstruction objective (binary cross-entropy on reconstructed methylation blocks and expression plus a KL term) with a cross-entropy classification term on the latent mean $\mu$. Two design choices make it feasible: methylation probes are split into 23 fully-connected blocks by target chromosome before being merged, which cuts the parameter count and captures intra-chromosome structure; and the classifier branch regularises the latent space so it must both reconstruct the sample and separate classes. The two-phase schedule—$\beta=0$ first, then $\beta=1$ with transferred encoder and decoder weights—is what lets the model first organise unlabelled samples and then tune that organisation to the classification task.

What would settle it

Re-run the 10-fold experiment with the autoencoder trained from scratch on each training fold alone, leaving the test fold out of both the unsupervised and supervised phases; if the 34-class average falls materially below 97.49% or below the single-omics 96.37%, the claimed multi-omics advantage is partly an artifact of pre-training on the test folds.

Watch

Extended reading notes

Core claim

The central claim is that task-oriented feature extraction and classification can live in one network: OmiVAE's encoder compresses 450,804 molecular features into a 128-dimensional latent Gaussian, the decoder reconstructs the input, and a classifier attached to the latent mean predicts tissue of origin and normality. Training happens in two phases—an unsupervised VAE phase, after which the latent space already shows separable clusters for organs, tumour subtypes, and normal tissue, and a supervised phase in which the encoder weights are transferred and the classifier loss is added. In the paper's evaluation this end-to-end model reaches 97.49% average accuracy among 33 tumour types and normal samples (34 classes) and 97.88% among tumour types alone (33 classes), outperforming prior pan-cancer classifiers built from gene expression alone. The same comparison shows multi-omics input beating single-omics input in every mode, which the paper reads as evidence that the latent vector combines complementary information from DNA methylation and gene expression.

Load-bearing premise

The accuracy numbers assume each test group is kept out of the entire training process, including the label-free pre-training step; if that step saw all samples before the test splits were made, the reported figures are optimistic.

Editorial extensions

If this is right

  • Combining gene expression and DNA methylation in one latent space improves classification over either omics alone, in both the unsupervised-embedding and end-to-end settings.
  • Because the unsupervised phase is a parametric encoder, a trained OmiVAE can embed new samples without re-running the whole dataset, unlike t-SNE.
  • The latent structure learned without labels separates tissues, tumour subtypes, and normal samples, so the model doubles as an exploratory tool for unannotated cohorts.
  • End-to-end training with the classifier branch beats feeding OmiVAE's unsupervised embeddings to an SVM, showing that task-oriented fine-tuning of the latent space contributes to the accuracy gain.
  • On the same pan-cancer task, the reported numbers exceed prior classifiers that used gene expression only.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A natural next test, which the paper does not run, is to ablate the classifier branch by keeping $\beta=0$ throughout and using the same latent dimension; that would isolate how much of the 97.49% comes from the classifier loss as opposed to the VAE's unsupervised geometry.
  • The chromosome-block encoding suggests a general recipe for other genomics assays with positional structure: group features by genomic location when feeding a VAE; this could be tried on chromatin-state or other genome-wide data.
  • If the multi-omics gain replicates on independent cohorts outside the paper's dataset, the latent vector might serve as a shared representation for downstream tasks such as survival or drug response, not just tissue classification.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 6 minor

Summary. The paper proposes OmiVAE, an end-to-end variational autoencoder combined with a classifier for pan-cancer classification from multi-omics data (gene expression and DNA methylation) on TCGA. The model is trained in two phases: an unsupervised phase with only the VAE reconstruction loss, followed by a supervised phase in which the encoder output feeds a classifier. The authors report 97.49% average accuracy for 34 classes (33 tumour types plus normal) and 97.88% for 33 tumour types after stratified 10-fold cross-validation, and claim that multi-omics input outperforms either single-omics input and that the model outperforms previously reported methods.

Significance. If the evaluation protocol is sound, the paper makes a useful empirical contribution: it demonstrates that a VAE-based latent representation can separate tumour types from multi-omics data, and it provides a public implementation. The explicit comparison between single-omics and multi-omics inputs is valuable, as is the quantitative evaluation of unsupervised embeddings against PCA, t-SNE, and UMAP. However, the central accuracy claims hinge on whether the unsupervised pretraining is properly nested inside the cross-validation folds, which the manuscript does not make clear. The cross-paper comparisons with prior methods are also weaker than the internal comparisons. These issues make the manuscript suitable for major revision rather than acceptance in its current form.

major comments (4)
  1. [Section IV-B] The text does not state that the unsupervised VAE pretraining is repeated inside each fold of the 10-fold cross-validation. It says only that the data were 'randomly divided into training, validation and testing sets' and that encoder/decoder parameters learned in the unsupervised phase were transferred to the supervised phase. If the pretraining is performed once on all 9,081 samples — as the description of Figure 2 suggests for the 2D embedding — then the test-fold feature distributions have already been seen by the encoder before supervised training begins, so the reported 97.49% and 97.88% accuracies are not standard held-out generalization estimates and are likely optimistically biased. Please specify clearly that the entire unsupervised-plus-supervised pipeline is nested within each fold, or re-run the evaluation with per-fold pretraining.
  2. [Section IV-A and Table I] The unsupervised-phase comparison suffers from the same ambiguity. The 2D embedding is described as learned from 'all 9,081 samples', and then an SVM is evaluated on the resulting representations with 10-fold cross-validation. If the dimensionality-reduction method (OmiVAE, PCA, KPCA, t-SNE, or UMAP) is fit on the full dataset before the SVM folds are created, the test representations are not independent of the training data. The dimensionality-reduction step must be fit on the training fold only for each of the 10 folds, and this must be stated explicitly for the reported Table I numbers to be valid.
  3. [Section IV-B, Tables II and III] The claim that OmiVAE 'outperformed other existing methods' is based on accuracies reported in references [16]-[18], which are gene-expression-only models evaluated on possibly different sample sets, preprocessing pipelines, and cross-validation schemes. These are not direct comparisons under a common protocol. To support the superiority claim, the baseline methods should be reimplemented and evaluated on the same preprocessed data (or at least the same expression-only subset) under identical cross-validation, or the claim should be softened to state that the reported accuracies are higher than previously reported figures.
  4. [Section IV-B, Tables II and III] The reported advantage of multi-omics over single-omics (97.49% versus 96.37% for 34 classes; 97.88% versus 96.86/96.89% for 33 classes) is not accompanied by any statistical significance test or confidence interval beyond the standard deviations across the 10 folds. Given the small margins, please add a paired significance test (for example, McNemar's test on the combined predictions or a paired test across folds) before concluding that complementary information from multiple omics types improves classification.
minor comments (6)
  1. [Abstract and Section I] The abstract and the introduction report 96.42% accuracy for gene-expression-only OmiVAE, while Table II reports 96.37%; please correct this inconsistency.
  2. [Section III-C, Equation (7)] The total loss is defined with weights α and β, and the text states that β is 0 in the unsupervised phase and 1 in the supervised phase, but the value of α is never reported. Please provide all hyperparameters, including α, the latent dimension, learning rate, batch size, and the early-stopping rule, for reproducibility.
  3. [Tables II and III] For the rows labeled 'OmiVAE+SVM (unsupervised phase)', the latent dimension is not specified; Table I uses a 2D bottleneck, but Section IV-B says the latent dimension is 'normally set to 128' in the supervised phase. Please state which latent dimension is used for each table.
  4. [Section III-C, Equation (4)] The VAE reconstruction loss uses binary cross-entropy for the log2-transformed and normalized gene-expression values, which are continuous rather than Bernoulli; please justify this choice or switch to a continuous loss such as mean squared error for the expression component.
  5. [Section IV] The text says the data were 'randomly divided into training, validation and testing sets' and also that stratified 10-fold cross-validation was used; please clarify how the validation set is formed within each fold and how early stopping is applied relative to training and testing.
  6. [General presentation] There are several typographical and LaTeX issues: 'In OmiV AE, we combines' in Section III-C, the GitHub URL rendered as 'OmiV AE' with a space, and 'Beat values' instead of 'Beta values' in Section III-B.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation found; the accuracy claims are empirical evaluation results, not outputs defined by inputs.

full rationale

The paper proposes OmiVAE and evaluates it against held-out labels via stratified 10-fold cross-validation. The unsupervised phase aims to learn a latent representation without labels; the supervised phase trains a classifier with labels. Neither the VAE objective (Eq. 2), the total loss (Eq. 7), nor the classification pipeline defines the reported accuracy in terms of the inputs used to fit the model. There are no self-citations used as load-bearing premises: references to previous pan-cancer classifiers (Li et al., Lyu and Haque, Mostavi et al.) are external benchmark comparisons, not justification for OmiVAE's architecture. The paper does not invoke uniqueness theorems or import an ansatz from prior work by the same authors. A legitimate experimental concern remains: Section IV-B says encoder/decoder parameters 'learned in the unsupervised phase were transferred to the supervised phase' and Figure 2 describes an unsupervised embedding of 'all 9,081 samples,' without explicitly stating that unsupervised pretraining is rerun inside each fold. If pretraining used the full dataset, the 10-fold accuracies could be optimistically biased; this would be a data-leakage or evaluation-validity issue, not a definitional or self-citational circularity. Under the circularity rubric, the derivation chain is self-contained: predictions are not fitted parameters renamed, and no output is an input by construction.

Assumptions & free parameters 6 free parameters · 6 assumptions · 0 invented entities

The central claim rests on standard VAE machinery, TCGA label accuracy, preprocessing choices (Y-chromosome removal, zero-expression filters, 10% missingness threshold, mean imputation), the chromosome-block inductive bias, and the assumption that 10-fold CV on TCGA generalizes. The model has several hand-set hyperparameters and no sensitivity analysis. No invented physical entities are introduced.

free parameters (6)
  • alpha (VAE loss weight) = not reported
    Equation (7) defines the total loss as alpha times the VAE loss plus beta times the classification loss, but alpha is never assigned a value; the reconstruction versus classification tradeoff is not audited.
  • beta (classification loss weight) = 0 in unsupervised phase, 1 in supervised phase
    beta is chosen by hand to switch between the two training phases, and this choice directly controls how much the classifier shapes the latent representation.
  • latent dimension = 128 for classification, 2 for visualization
    The bottleneck size is set manually; it controls how much information is preserved and therefore affects classification accuracy.
  • hidden layer sizes = 1,024 and 4,096 wide layers; 23 chromosome blocks of 256
    Architecture widths are chosen by hand to balance capacity and GPU memory; no sensitivity analysis is provided.
  • learning rate and batch size = 1e-3 and 32
    Standard Adam settings are used, with early stopping, but the early stopping criterion is not specified.
  • missing value cutoff for feature filtering = 10%
    Exons and CpG probes with N/A in more than 10% of samples were removed (Section III-B); changing this threshold changes the input feature set and therefore the accuracy estimates.
assumptions (6)
  • standard math The VAE evidence lower bound and reparameterization are valid.
    The model relies on the standard VAE objective and Gaussian reparameterization without proof, which is accepted background in the literature.
  • domain assumption TCGA sample labels are accurate tissue-of-origin and tumor/normal annotations.
    Classification accuracy is measured against TCGA labels; label errors would directly change the reported numbers.
  • domain assumption Preprocessing filters and mean imputation preserve the signal needed for classification.
    Removing Y-chromosome features, zero-expression exons, high-missingness probes, and imputing remaining missing values are choices that could discard or distort signal; no sensitivity analysis is given.
  • domain assumption Pooling all normal samples into one class is a meaningful 34-class setup.
    Normal samples come from many tissues and are genetically distinct by tissue; the model may be recognizing tissue-of-origin rather than cancer status, and the paper does not discuss this confound.
  • ad hoc to paper Chromosome-block encoding is a valid inductive bias for methylation data.
    Splitting CpG probes by chromosome into separate fully connected blocks reduces parameters but assumes intra-chromosome relationships dominate; this assumption is unproven.
  • domain assumption 10-fold cross-validation on TCGA estimates real-world generalization.
    No external cohort is used, and the ambiguity about whether pre-training is nested inside folds makes the cross-validation estimate uncertain.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Integrated Multi-omics Analysis Using Variational Autoencoders: Application to Pan-cancer Classification." pith.science (2026). https://pith.science/paper/VZZTB3GI

@misc{pith2026190806278,
  author       = {Pith},
  title        = {Pith review of: Integrated Multi-omics Analysis Using Variational Autoencoders: Application to Pan-cancer Classification},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VZZTB3GI}},
  note         = {Machine review of arXiv:1908.06278}
}
read the original abstract

Different aspects of a clinical sample can be revealed by multiple types of omics data. Integrated analysis of multi-omics data provides a comprehensive view of patients, which has the potential to facilitate more accurate clinical decision making. However, omics data are normally high dimensional with large number of molecular features and relatively small number of available samples with clinical labels. The "dimensionality curse" makes it challenging to train a machine learning model using high dimensional omics data like DNA methylation and gene expression profiles. Here we propose an end-to-end deep learning model called OmiVAE to extract low dimensional features and classify samples from multi-omics data. OmiVAE combines the basic structure of variational autoencoders with a classification network to achieve task-oriented feature extraction and multi-class classification. The training procedure of OmiVAE is comprised of an unsupervised phase without the classifier and a supervised phase with the classifier. During the unsupervised phase, a hierarchical cluster structure of samples can be automatically formed without the need for labels. And in the supervised phase, OmiVAE achieved an average classification accuracy of 97.49% after 10-fold cross-validation among 33 tumour types and normal samples, which shows better performance than other existing methods. The OmiVAE model learned from multi-omics data outperformed that using only one type of omics data, which indicates that the complementary information from different omics datatypes provides useful insights for biomedical tasks like cancer classification.

Figures

Figures reproduced from arXiv: 1908.06278 by the authors.

Figure 1
Figure 1. Diagram showing the architecture of OmiVAE. The network structure [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. 2D embedding of the multi-omics data learned by the unsupervised phase of OmiVAE (left) and PCA (right). Samples of various tumour types were [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Multi-omics latent vectors of tumours and corresponding normal [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Visualisation of the classification space learned by the end-to-end [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

24 extracted references · 18 canonical work pages

  1. [16]

    A comprehensive genomic pan-cancer classification using the cancer genome atlas gene expression data,

    Y . Li, K. Kang, J. M. Krahn, N. Croutwater, K. Lee, D. M. Umbach, and L. Li, “A comprehensive genomic pan-cancer classification using the cancer genome atlas gene expression data,” BMC Genomics, vol. 18, no. 1, p. 508, 2017

    work page 2017

  2. [18]

    Convolutional neural network models for cancer type prediction based on gene expression

    M. Mostavi, Y .-C. Chiu, Y . Huang, and Y . Chen, “Convolutional neural network models for cancer type prediction based on gene expression,” arXiv preprint arXiv:1906.07794 , 2019

    work page Pith review arXiv 1906

  3. [1]

    Computational solutions for omics data,

    B. Berger, J. Peng, and M. Singh, “Computational solutions for omics data,” Nature Reviews Genetics , vol. 14, no. 5, p. 333, 2013

    work page 2013

  4. [2]

    Pam50 assay and the three- gene model for identifying the major and clinically relevant molecular subtypes of breast cancer,

    A. Prat, J. Parker, C. Fan, and C. Perou, “Pam50 assay and the three- gene model for identifying the major and clinically relevant molecular subtypes of breast cancer,” Breast Cancer Research and Treatment, vol. 135, no. 1, pp. 301–306, 2012

    work page 2012

  5. [3]

    Deep learning,

    Y . LeCun, Y . Bengio, and G. Hinton, “Deep learning,”Nature, vol. 521, no. 7553, p. 436, 2015

    2015

  6. [4]

    Auto-encoding variational bayes,

    D. P. Kingma and M. Welling, “Auto-encoding variational bayes,” arXiv preprint arXiv:1312.6114, 2013

    arXiv 2013

  7. [5]

    The cancer genome atlas pan-cancer analysis project,

    J. N. Weinstein, E. A. Collisson, G. B. Mills, K. R. M. Shaw, B. A. Ozenberger, K. Ellrott, I. Shmulevich, C. Sander, J. M. Stuart, C. G. A. R. Network et al. , “The cancer genome atlas pan-cancer analysis project,” Nature Genetics, vol. 45, no. 10, p. 1113, 2013

    work page 2013

  8. [6]

    Deep learning– based multi-omics integration robustly predicts survival in liver cancer,

    K. Chaudhary, O. B. Poirion, L. Lu, and L. X. Garmire, “Deep learning– based multi-omics integration robustly predicts survival in liver cancer,” Clinical Cancer Research , vol. 24, no. 6, pp. 1248–1259, 2018

    work page 2018

Show all 24 references
  1. [7]

    Unsupervised classification of multi-omics data during cardiac remodeling using deep learning,

    N. C. Chung, B. Mirza, H. Choi, J. Wang, D. Wang, P. Ping, and W. Wang, “Unsupervised classification of multi-omics data during cardiac remodeling using deep learning,” Methods, 2019

    2019

  2. [8]

    Multi-view factorization autoencoder with net- work constraints for multi-omic integrative analysis,

    T. Ma and A. Zhang, “Multi-view factorization autoencoder with net- work constraints for multi-omic integrative analysis,” in 2018 IEEE International Conference on Bioinformatics and Biomedicine (BIBM) , 2018, pp. 702–707

    2018

  3. [9]

    Using deep learning to enhance cancer diagnosis and classification,

    R. Fakoor, F. Ladhak, A. Nazi, and M. Huber, “Using deep learning to enhance cancer diagnosis and classification,” in Proceedings of the International Conference on Machine Learning (ICML-13) , vol. 28, 2013

    2013

  4. [10]

    A deep learning approach for cancer detection and relevant gene identification,

    P. Danaee, R. Ghaeini, and D. A. Hendrix, “A deep learning approach for cancer detection and relevant gene identification,” in Pacific Symposium on Biocomputing 2017 , 2017, pp. 219–229

    2017

  5. [11]

    Extracting a biologically relevant latent space from cancer transcriptomes with variational autoencoders,

    G. P. Way and C. S. Greene, “Extracting a biologically relevant latent space from cancer transcriptomes with variational autoencoders,” in Pacific Symposium on Biocomputing 2018 , 2018, pp. 80–91

    2018

  6. [12]

    A new dimension of breast cancer epigenetics,

    A. J. Titus, C. A. Bobak, and B. C. Christensen, “A new dimension of breast cancer epigenetics,” in 9th International Conference on Bioinfor- matics Models, Methods and Algorithms , 2018

    2018

  7. [13]

    Exploring dna methylation data of lung cancer samples with variational autoencoders,

    Z. Wang and Y . Wang, “Exploring dna methylation data of lung cancer samples with variational autoencoders,” in 2018 IEEE International Conference on Bioinformatics and Biomedicine (BIBM), 2018, pp. 1286– 1289

    2018

  8. [14]

    Hybrid approach of relation network and localized graph convolutional filtering for breast cancer subtype classification,

    S. Rhee, S. Seo, and S. Kim, “Hybrid approach of relation network and localized graph convolutional filtering for breast cancer subtype classification,” in Proceedings of the Twenty-Seventh International Joint Conference on Artificial Intelligence, IJCAI-18 , 2018, pp. 3527–3534

    2018

  9. [15]

    Dna methylation- based classification of central nervous system tumours,

    D. Capper, D. T. Jones, M. Sill, V . Hovestadt, D. Schrimpf, D. Sturm, C. Koelsche, F. Sahm, L. Chavez, D. E. Reuss et al., “Dna methylation- based classification of central nervous system tumours,” Nature, vol. 555, no. 7697, p. 469, 2018

    2018

  10. [17]

    Deep learning based tumor type classification using gene expression data,

    B. Lyu and A. Haque, “Deep learning based tumor type classification using gene expression data,” in Proceedings of the 2018 ACM Inter- national Conference on Bioinformatics, Computational Biology, and Health Informatics, 2018, pp. 89–96

    2018

  11. [19]

    Goodfellow, Y

    I. Goodfellow, Y . Bengio, and A. Courville, Deep learning. MIT press, 2016

    2016

  12. [20]

    Batch normalization: Accelerating deep network training by reducing internal covariate shift,

    S. Ioffe and C. Szegedy, “Batch normalization: Accelerating deep network training by reducing internal covariate shift,” arXiv preprint arXiv:1502.03167, 2015

    2015 arXiv

  13. [21]

    Rectified linear units improve restricted boltzmann machines,

    V . Nair and G. E. Hinton, “Rectified linear units improve restricted boltzmann machines,” in Proceedings of the International Conference on Machine Learning (ICML-10) , 2010, pp. 807–814

    2010

  14. [22]

    Adam: A method for stochastic optimization,

    D. P. Kingma and J. Ba, “Adam: A method for stochastic optimization,” International Conference on Learning Representations (ICLR 2015) , 2015

    2015

  15. [23]

    Visualizing data using t-sne,

    L. v. d. Maaten and G. Hinton, “Visualizing data using t-sne,” Journal of Machine Learning Research , vol. 9, no. Nov, pp. 2579–2605, 2008

    2008

  16. [24]

    Umap: Uniform manifold approximation and projection for dimension reduction,

    L. McInnes, J. Healy, and J. Melville, “Umap: Uniform manifold approximation and projection for dimension reduction,” arXiv preprint arXiv:1802.03426, 2018

    2018 arXiv

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.