Abstract
Aims
The aim of this work is to do an integrated multi-omics analysis of ovarian cancer using deep learning (DL), namely
variational autoencoder (VAE) and an improved version of VAE named Maximum Mean Discrepancy VAE (MMD-VAE).
Background
Cancer is a complex disease that deregulates cellular functions at various molecular levels. 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, 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, which makes a DL based integrated multi-omics analysis difficult. DL-based dimensionality reduction
technique, including 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 only.
Methods
We used mono-omics and multi-omics (i.e., di- and tri-omics) data for the study. First, we have downloaded four different
mono-omics TCGA datasets of ovarian cancer and generated multi-omics data using different combinations of these
mono-omics data. Then, we designed and developed a DL architecture for VAE and MMD-VAE and used the architecture
to learn latent and compressed features from the mono-omics, di-omics and tri-omics datasets. Finally, we used the
learned latent compressed features to analyse ovarian cancer through cancer samples identification, molecular subtypes
clustering and classification.
Results
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\%.
Conclusions
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.
The aim of this work is to do an integrated multi-omics analysis of ovarian cancer using deep learning (DL), namely
variational autoencoder (VAE) and an improved version of VAE named Maximum Mean Discrepancy VAE (MMD-VAE).
Background
Cancer is a complex disease that deregulates cellular functions at various molecular levels. 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, 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, which makes a DL based integrated multi-omics analysis difficult. DL-based dimensionality reduction
technique, including 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 only.
Methods
We used mono-omics and multi-omics (i.e., di- and tri-omics) data for the study. First, we have downloaded four different
mono-omics TCGA datasets of ovarian cancer and generated multi-omics data using different combinations of these
mono-omics data. Then, we designed and developed a DL architecture for VAE and MMD-VAE and used the architecture
to learn latent and compressed features from the mono-omics, di-omics and tri-omics datasets. Finally, we used the
learned latent compressed features to analyse ovarian cancer through cancer samples identification, molecular subtypes
clustering and classification.
Results
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\%.
Conclusions
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.
| Original language | English |
|---|---|
| Pages | 149 |
| Number of pages | 1 |
| Publication status | Published - May 2021 |
| Event | British Gynaecological Cancer Society Annual Scientific Meeting 2021 - Duration: 13 May 2021 → 14 May 2021 |
Conference
| Conference | British Gynaecological Cancer Society Annual Scientific Meeting 2021 |
|---|---|
| Period | 13/05/21 → 14/05/21 |