Unsupervised Machine Learning for Data Encoding applied to Ovarian Cancer Transcriptomes

George, Tom M.; Lió, Píetro

doi:10.1101/855593

Cited by 5 publications

(7 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…There are many works which have used VAE in their studies. However, they are mostly mono-omics studies of individual cancer 28,32,33 or pancancer 29,34,35 . OmiVAE 29 is the only work that considered VAE for integrated multi-omics (di-omics) analysis of pancancer.…”

mentioning

confidence: 99%

See 1 more Smart Citation

Integrated multi-omics analysis of ovarian cancer using variational autoencoders

Hira

Razzaque

Angione

et al. 2021

Sci Rep

View full text Add to dashboard Cite

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.

show abstract

mentioning

confidence: 99%

“…Moreover, most of the existing works 28,32,33,35 use unsupervised dimensionality reduction methods, separating the downstream analysis from the reduction method. However, dimensionality reduction in cancer multi-omics analysis is an intermediate step toward the downstream analysis, like classification (e.g., cancer vs normal cell).…”

mentioning

confidence: 99%

Integrated multi-omics analysis of ovarian cancer using variational autoencoders

Hira

Razzaque

Angione

et al. 2021

Sci Rep

View full text Add to dashboard Cite

show abstract

“…Multiple studies (An and Cho, 2015; Li and She, 2017; Bouchacourt et al , 2017; Kipf and Welling, 2016) have established the power of the variational autoencoder (VAE; Kingma and Welling (2013); Jimenez Rezende et al (2014))—an unsupervised nonlinear data embedding model with two deep neural networks oppositely connected through a low-dimensional probabilistic latent space—for finding meaningful and useful latent features in high-dimensional data. In the context of cancer bioinformatics, VAEs have been variously used to (i) model cancer gene expression and capture biologically-relevant features using the TCGA Pan-cancer Project RNA-seq dataset (Way and Greene, 2018); (ii) find encodings that correlate with biological features such as patient sex and tumor type (Titus et al , 2018); (iii) find encodings that can be used to predict gene inactivation in cancer (Way and Greene, 2017); and (iv) obtain an encoding that is predictive of chemotherapy resistance (George and Lio, 2019). Based on their exploration of multiple VAE architectures for predicting gene inactivation in a pan-cancer dataset, Way & Greene reported (2017) biological insights obtained from the latent-space embeddings learned by VAEs.…”

Section: Introductionmentioning

confidence: 99%

“…Based on their exploration of multiple VAE architectures for predicting gene inactivation in a pan-cancer dataset, Way & Greene reported (2017) biological insights obtained from the latent-space embeddings learned by VAEs. George and Lio (2019) used a VAE-based, fully unsupervised approach to encode ovarian tumor transcriptomes and obtained latent-space features that were associated with response to chemotherapy. These studies suggest that a tumor transcriptome VAE may be broadly useful for the response-to-chemotherapy prediction problem and they set the stage for the present multi-cancer investigation of the utility of the tumor transcriptome VAE in precision oncology.…”

Section: Introductionmentioning

confidence: 99%

Predicting chemotherapy response using a variational autoencoder approach

Wei

Ramsey

2021

Preprint

View full text Add to dashboard Cite

MotivationMultiple studies have shown the utility of transcriptome-wide RNA-seq profiles as features for machine learning-based prediction of response to chemotherapy in cancer. While tumor transcriptome profiles are publicly available for thousands of tumors for many cancer types, a relatively modest number of tumor profiles are clinically annotated for response to chemotherapy. The paucity of labeled examples and high dimension of the feature data limit performance for predicting therapeutic response using fully-supervised classification methods. Recently, multiple studies have established the utility of a deep neural network approach, the variational autoencoder (VAE), for generating meaningful latent features from original data. Here, we report first study of a semi-supervised approach using VAE-encoded tumor transcriptome features and regularized gradient boosted decision trees (XGBoost) to predict chemotherapy drug response for five cancer types: colon adenocarcinoma, pancreatic adenocarcinoma, bladder carcinoma, sarcoma, and breast invasive carcinoma.ResultsWe found: (1) VAE-encoding of the tumor transcriptome preserves the cancer type identity of the tumor, suggesting preservation of biologically relevant information; and (2) as a feature-set for supervised classification to predict response-to-chemotherapy, the unsupervised VAE encoding of the tumor’s gene expression profile leads to better area under the receiver operating characteristic curve (AUROC) classification performance than either the original gene expression profile or the PCA principal components of the gene expression profile, in four out of five cancer types that we tested.Availabilitygithub.com/ATHED/VAE_for_chemotherapy_drug_response_predictionContactramseyst@oregonstate.eduSupplementary informationSupplementary data are available at Bioinformatics online.

show abstract

“…Way and Greene [ 28 ] explored VAE architectures for predicting gene inactivation in a pan-cancer dataset and reported biological insights obtained from the latent-space embeddings. George and Lio [ 29 ] used a VAE-based, unsupervised approach to encode tumor transcriptomes to obtain latent-space features associated with chemotherapy response. Dincer et al [ 30 ] used a semi-supervised, VAE-lasso approach to predict drug sensitivity of cancer cells in vitro.…”

Section: Introductionmentioning

confidence: 99%

Predicting chemotherapy response using a variational autoencoder approach

Wei

Ramsey

2021

BMC Bioinformatics

View full text Add to dashboard Cite

Background Multiple studies have shown the utility of transcriptome-wide RNA-seq profiles as features for machine learning-based prediction of response to chemotherapy in cancer. While tumor transcriptome profiles are publicly available for thousands of tumors for many cancer types, a relatively modest number of tumor profiles are clinically annotated for response to chemotherapy. The paucity of labeled examples and the high dimension of the feature data limit performance for predicting therapeutic response using fully-supervised classification methods. Recently, multiple studies have established the utility of a deep neural network approach, the variational autoencoder (VAE), for generating meaningful latent features from original data. Here, we report the first study of a semi-supervised approach using VAE-encoded tumor transcriptome features and regularized gradient boosted decision trees (XGBoost) to predict chemotherapy drug response for five cancer types: colon, pancreatic, bladder, breast, and sarcoma. Results We found: (1) VAE-encoding of the tumor transcriptome preserves the cancer type identity of the tumor, suggesting preservation of biologically relevant information; and (2) as a feature-set for supervised classification to predict response-to-chemotherapy, the unsupervised VAE encoding of the tumor’s gene expression profile leads to better area under the receiver operating characteristic curve and area under the precision-recall curve classification performance than the original gene expression profile or the PCA principal components or the ICA components of the gene expression profile, in four out of five cancer types that we tested. Conclusions Given high-dimensional “omics” data, the VAE is a powerful tool for obtaining a nonlinear low-dimensional embedding; it yields features that retain biological patterns that distinguish between different types of cancer and that enable more accurate tumor transcriptome-based prediction of response to chemotherapy than would be possible using the original data or their principal components.

show abstract

Unsupervised Machine Learning for Data Encoding applied to Ovarian Cancer Transcriptomes

Cited by 5 publications

References 30 publications

Integrated multi-omics analysis of ovarian cancer using variational autoencoders

Integrated multi-omics analysis of ovarian cancer using variational autoencoders

Predicting chemotherapy response using a variational autoencoder approach

Predicting chemotherapy response using a variational autoencoder approach

Contact Info

Product

Resources

About