Deconvolution of autoencoders to learn biological regulatory modules from single cell mRNA sequencing data

Interpretable deep learning in single-cell omics

MOTIVATION

Single-cell omics technologies have enabled the quantification of molecular profiles in individual cells at an unparalleled resolution. Deep learning, a rapidly evolving sub-field of machine learning, has instilled a significant interest in single-cell omics research due to its remarkable success in analysing heterogeneous high-dimensional single-cell omics data. Nevertheless, the inherent multi-layer nonlinear architecture of deep learning models often makes them 'black boxes' as the reasoning behind predictions is often unknown and not transparent to the user. This has stimulated an increasing body of research for addressing the lack of interpretability in deep learning models, especially in single-cell omics data analyses, where the identification and understanding of molecular regulators are crucial for interpreting model predictions and directing downstream experimental validations.

RESULTS

In this work, we introduce the basics of single-cell omics technologies and the concept of interpretable deep learning. This is followed by a review of the recent interpretable deep learning models applied to various single-cell omics research. Lastly, we highlight the current limitations and discuss potential future directions.

Model-based evaluation of spatiotemporal data reduction methods with unknown ground truth through optimal visualization and interpretability metrics

Optimizing and benchmarking data reduction methods for dynamic or spatial visualization and interpretation (DSVI) face challenges due to many factors, including data complexity, lack of ground truth, time-dependent metrics, dimensionality bias and different visual mappings of the same data. Current studies often focus on independent static visualization or interpretability metrics that require ground truth. To overcome this limitation, we propose the MIBCOVIS framework, a comprehensive and interpretable benchmarking and computational approach. MIBCOVIS enhances the visualization and interpretability of high-dimensional data without relying on ground truth by integrating five robust metrics, including a novel time-ordered Markov-based structural metric, into a semi-supervised hierarchical Bayesian model. The framework assesses method accuracy and considers interaction effects among metric features. We apply MIBCOVIS using linear and nonlinear dimensionality reduction methods to evaluate optimal DSVI for four distinct dynamic and spatial biological processes captured by three single-cell data modalities: CyTOF, scRNA-seq and CODEX. These data vary in complexity based on feature dimensionality, unknown cell types and dynamic or spatial differences. Unlike traditional single-summary score approaches, MIBCOVIS compares accuracy distributions across methods. Our findings underscore the joint evaluation of visualization and interpretability, rather than relying on separate metrics. We reveal that prioritizing average performance can obscure method feature performance. Additionally, we explore the impact of data complexity on visualization and interpretability. Specifically, we provide optimal parameters and features and recommend methods, like the optimized variational contractive autoencoder, for targeted DSVI for various data complexities. MIBCOVIS shows promise for evaluating dynamic single-cell atlases and spatiotemporal data reduction models.

NCAE: data-driven representations using a deep network-coherent DNA methylation autoencoder identify robust disease and risk factor signatures

Precision medicine relies on the identification of robust disease and risk factor signatures from omics data. However, current knowledge-driven approaches may overlook novel or unexpected phenomena due to the inherent biases in biological knowledge. In this study, we present a data-driven signature discovery workflow for DNA methylation analysis utilizing network-coherent autoencoders (NCAEs) with biologically relevant latent embeddings. First, we explored the architecture space of autoencoders trained on a large-scale pan-tissue compendium (n = 75 272) of human epigenome-wide association studies. We observed the emergence of co-localized patterns in the deep autoencoder latent space representations that corresponded to biological network modules. We determined the NCAE configuration with the strongest co-localization and centrality signals in the human protein interactome. Leveraging the NCAE embeddings, we then trained interpretable deep neural networks for risk factor (aging, smoking) and disease (systemic lupus erythematosus) prediction and classification tasks. Remarkably, our NCAE embedding-based models outperformed existing predictors, revealing novel DNA methylation signatures enriched in gene sets and pathways associated with the studied condition in each case. Our data-driven biomarker discovery workflow provides a generally applicable pipeline to capture relevant risk factor and disease information. By surpassing the limitations of knowledge-driven methods, our approach enhances the understanding of complex epigenetic processes, facilitating the development of more effective diagnostic and therapeutic strategies.

Handling missing values in healthcare data: A systematic review of deep learning-based imputation techniques

OBJECTIVE

The proper handling of missing values is critical to delivering reliable estimates and decisions, especially in high-stakes fields such as clinical research. In response to the increasing diversity and complexity of data, many researchers have developed deep learning (DL)-based imputation techniques. We conducted a systematic review to evaluate the use of these techniques, with a particular focus on the types of data, intending to assist healthcare researchers from various disciplines in dealing with missing data.

MATERIALS AND METHODS

We searched five databases (MEDLINE, Web of Science, Embase, CINAHL, and Scopus) for articles published prior to February 8, 2023 that described the use of DL-based models for imputation. We examined selected articles from four perspectives: data types, model backbones (i.e., main architectures), imputation strategies, and comparisons with non-DL-based methods. Based on data types, we created an evidence map to illustrate the adoption of DL models.

RESULTS

Out of 1822 articles, a total of 111 were included, of which tabular static data (29%, 32/111) and temporal data (40%, 44/111) were the most frequently investigated. Our findings revealed a discernible pattern in the choice of model backbones and data types, for example, the dominance of autoencoder and recurrent neural networks for tabular temporal data. The discrepancy in imputation strategy usage among data types was also observed. The "integrated" imputation strategy, which solves the imputation task simultaneously with downstream tasks, was most popular for tabular temporal data (52%, 23/44) and multi-modal data (56%, 5/9). Moreover, DL-based imputation methods yielded a higher level of imputation accuracy than non-DL methods in most studies.

CONCLUSION

The DL-based imputation models are a family of techniques, with diverse network structures. Their designation in healthcare is usually tailored to data types with different characteristics. Although DL-based imputation models may not be superior to conventional approaches across all datasets, it is highly possible for them to achieve satisfactory results for a particular data type or dataset. There are, however, still issues with regard to portability, interpretability, and fairness associated with current DL-based imputation models.

A systematic review of biologically-informed deep learning models for cancer: fundamental trends for encoding and interpreting oncology data

BACKGROUND

There is an increasing interest in the use of Deep Learning (DL) based methods as a supporting analytical framework in oncology. However, most direct applications of DL will deliver models with limited transparency and explainability, which constrain their deployment in biomedical settings.

METHODS

This systematic review discusses DL models used to support inference in cancer biology with a particular emphasis on multi-omics analysis. It focuses on how existing models address the need for better dialogue with prior knowledge, biological plausibility and interpretability, fundamental properties in the biomedical domain. For this, we retrieved and analyzed 42 studies focusing on emerging architectural and methodological advances, the encoding of biological domain knowledge and the integration of explainability methods.

RESULTS

We discuss the recent evolutionary arch of DL models in the direction of integrating prior biological relational and network knowledge to support better generalisation (e.g. pathways or Protein-Protein-Interaction networks) and interpretability. This represents a fundamental functional shift towards models which can integrate mechanistic and statistical inference aspects. We introduce a concept of bio-centric interpretability and according to its taxonomy, we discuss representational methodologies for the integration of domain prior knowledge in such models.

CONCLUSIONS

The paper provides a critical outlook into contemporary methods for explainability and interpretability used in DL for cancer. The analysis points in the direction of a convergence between encoding prior knowledge and improved interpretability. We introduce bio-centric interpretability which is an important step towards formalisation of biological interpretability of DL models and developing methods that are less problem- or application-specific.

clusterMaker2: a major update to clusterMaker, a multi-algorithm clustering app for Cytoscape

BACKGROUND

Since the initial publication of clusterMaker, the need for tools to analyze large biological datasets has only increased. New datasets are significantly larger than a decade ago, and new experimental techniques such as single-cell transcriptomics continue to drive the need for clustering or classification techniques to focus on portions of datasets of interest. While many libraries and packages exist that implement various algorithms, there remains the need for clustering packages that are easy to use, integrated with visualization of the results, and integrated with other commonly used tools for biological data analysis. clusterMaker2 has added several new algorithms, including two entirely new categories of analyses: node ranking and dimensionality reduction. Furthermore, many of the new algorithms have been implemented using the Cytoscape jobs API, which provides a mechanism for executing remote jobs from within Cytoscape. Together, these advances facilitate meaningful analyses of modern biological datasets despite their ever-increasing size and complexity.

RESULTS

The use of clusterMaker2 is exemplified by reanalyzing the yeast heat shock expression experiment that was included in our original paper; however, here we explored this dataset in significantly more detail. Combining this dataset with the yeast protein-protein interaction network from STRING, we were able to perform a variety of analyses and visualizations from within clusterMaker2, including Leiden clustering to break the entire network into smaller clusters, hierarchical clustering to look at the overall expression dataset, dimensionality reduction using UMAP to find correlations between our hierarchical visualization and the UMAP plot, fuzzy clustering, and cluster ranking. Using these techniques, we were able to explore the highest-ranking cluster and determine that it represents a strong contender for proteins working together in response to heat shock. We found a series of clusters that, when re-explored as fuzzy clusters, provide a better presentation of mitochondrial processes.

CONCLUSIONS

clusterMaker2 represents a significant advance over the previously published version, and most importantly, provides an easy-to-use tool to perform clustering and to visualize clusters within the Cytoscape network context. The new algorithms should be welcome to the large population of Cytoscape users, particularly the new dimensionality reduction and fuzzy clustering techniques.

siVAE: interpretable deep generative models for single-cell transcriptomes

Neural networks such as variational autoencoders (VAE) perform dimensionality reduction for the visualization and analysis of genomic data, but are limited in their interpretability: it is unknown which data features are represented by each embedding dimension. We present siVAE, a VAE that is interpretable by design, thereby enhancing downstream analysis tasks. Through interpretation, siVAE also identifies gene modules and hubs without explicit gene network inference. We use siVAE to identify gene modules whose connectivity is associated with diverse phenotypes such as iPSC neuronal differentiation efficiency and dementia, showcasing the wide applicability of interpretable generative models for genomic data analysis.

Interpretable generative deep learning: an illustration with single cell gene expression data

Deep generative models can learn the underlying structure, such as pathways or gene programs, from omics data. We provide an introduction as well as an overview of such techniques, specifically illustrating their use with single-cell gene expression data. For example, the low dimensional latent representations offered by various approaches, such as variational auto-encoders, are useful to get a better understanding of the relations between observed gene expressions and experimental factors or phenotypes. Furthermore, by providing a generative model for the latent and observed variables, deep generative models can generate synthetic observations, which allow us to assess the uncertainty in the learned representations. While deep generative models are useful to learn the structure of high-dimensional omics data by efficiently capturing non-linear dependencies between genes, they are sometimes difficult to interpret due to their neural network building blocks. More precisely, to understand the relationship between learned latent variables and observed variables, e.g., gene transcript abundances and external phenotypes, is difficult. Therefore, we also illustrate current approaches that allow us to infer the relationship between learned latent variables and observed variables as well as external phenotypes. Thereby, we render deep learning approaches more interpretable. In an application with single-cell gene expression data, we demonstrate the utility of the discussed methods.

Application of deep learning methods: From molecular modelling to patient classification

We are now well into the information driven age with complex, heterogeneous, datasets in the biological sciences continuing to grow at a rapid pace. Moreover, distilling of such datasets, to find new governing principles, are underway. Leading the surge are new and exciting algorithmic developments in computer simulation and machine learning, most notably for the latter, those centred on deep learning. However, practical applications of cell centric computations within the biological sciences, even when carefully benchmarked against existing experimental datasets, remain challenging. Here we discuss the application of deep learning methodologies to support our understanding of cell functionality and as an aid to patient classification. Whilst comprehensive end-to-end deep learning approaches that utilise knowledge of the cell and its molecular components to aid human disease classification are yet to be implemented, important for opening the door to more effective molecular and cell-based therapies, we illustrate that many deep learning applications have been developed to tackle components of such an ambitious pipeline. We end our discussion on what the future may hold, especially how an integrated framework of computer simulations and deep learning, in conjunction with wet-bench experimentation, could enable to reveal the governing principles underlying cell functionalities within the tissue environments cells operate.

Large-Scale Integrative Analysis of Soybean Transcriptome Using an Unsupervised Autoencoder Model

Plant tissues are distinguished by their gene expression patterns, which can help identify tissue-specific highly expressed genes and their differential functional modules. For this purpose, large-scale soybean transcriptome samples were collected and processed starting from raw sequencing reads in a uniform analysis pipeline. To address the gene expression heterogeneity in different tissues, we utilized an adversarial deconfounding autoencoder (AD-AE) model to map gene expressions into a latent space and adapted a standard unsupervised autoencoder (AE) model to help effectively extract meaningful biological signals from the noisy data. As a result, four groups of 1,743, 914, 2,107, and 1,451 genes were found highly expressed specifically in leaf, root, seed and nodule tissues, respectively. To obtain key transcription factors (TFs), hub genes and their functional modules in each tissue, we constructed tissue-specific gene regulatory networks (GRNs), and differential correlation networks by using corrected and compressed gene expression data. We validated our results from the literature and gene enrichment analysis, which confirmed many identified tissue-specific genes. Our study represents the largest gene expression analysis in soybean tissues to date. It provides valuable targets for tissue-specific research and helps uncover broader biological patterns. Code is publicly available with open source at https://github.com/LingtaoSu/SoyMeta.

Interpretable Autoencoders Trained on Single Cell Sequencing Data Can Transfer Directly to Data from Unseen Tissues

Autoencoders have been used to model single-cell mRNA-sequencing data with the purpose of denoising, visualization, data simulation, and dimensionality reduction. We, and others, have shown that autoencoders can be explainable models and interpreted in terms of biology. Here, we show that such autoencoders can generalize to the extent that they can transfer directly without additional training. In practice, we can extract biological modules, denoise, and classify data correctly from an autoencoder that was trained on a different dataset and with different cells (a foreign model). We deconvoluted the biological signal encoded in the bottleneck layer of scRNA-models using saliency maps and mapped salient features to biological pathways. Biological concepts could be associated with specific nodes and interpreted in relation to biological pathways. Even in this unsupervised framework, with no prior information about cell types or labels, the specific biological pathways deduced from the model were in line with findings in previous research. It was hypothesized that autoencoders could learn and represent meaningful biology; here, we show with a systematic experiment that this is true and even transcends the training data. This means that carefully trained autoencoders can be used to assist the interpretation of new unseen data.

Feature selection revisited in the single-cell era

Recent advances in single-cell biotechnologies have resulted in high-dimensional datasets with increased complexity, making feature selection an essential technique for single-cell data analysis. Here, we revisit feature selection techniques and summarise recent developments. We review their application to a range of single-cell data types generated from traditional cytometry and imaging technologies and the latest array of single-cell omics technologies. We highlight some of the challenges and future directions and finally consider their scalability and make general recommendations on each type of feature selection method. We hope this review stimulates future research and application of feature selection in the single-cell era.

Improving gene function predictions using independent transcriptional components

The interpretation of high throughput sequencing data is limited by our incomplete functional understanding of coding and non-coding transcripts. Reliably predicting the function of such transcripts can overcome this limitation. Here we report the use of a consensus independent component analysis and guilt-by-association approach to predict over 23,000 functional groups comprised of over 55,000 coding and non-coding transcripts using publicly available transcriptomic profiles. We show that, compared to using Principal Component Analysis, Independent Component Analysis-derived transcriptional components enable more confident functionality predictions, improve predictions when new members are added to the gene sets, and are less affected by gene multi-functionality. Predictions generated using human or mouse transcriptomic data are made available for exploration in a publicly available web portal.

scVAE: variational auto-encoders for single-cell gene expression data

MOTIVATION

Models for analysing and making relevant biological inferences from massive amounts of complex single-cell transcriptomic data typically require several individual data-processing steps, each with their own set of hyperparameter choices. With deep generative models one can work directly with count data, make likelihood-based model comparison, learn a latent representation of the cells and capture more of the variability in different cell populations.

RESULTS

We propose a novel method based on variational auto-encoders (VAEs) for analysis of single-cell RNA sequencing (scRNA-seq) data. It avoids data preprocessing by using raw count data as input and can robustly estimate the expected gene expression levels and a latent representation for each cell. We tested several count likelihood functions and a variant of the VAE that has a priori clustering in the latent space. We show for several scRNA-seq datasets that our method outperforms recently proposed scRNA-seq methods in clustering cells and that the resulting clusters reflect cell types.

AVAILABILITY AND IMPLEMENTATION

Our method, called scVAE, is implemented in Python using the TensorFlow machine-learning library, and it is freely available at https://github.com/scvae/scvae.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Potential applications of deep learning in single-cell RNA sequencing analysis for cell therapy and regenerative medicine

When used in cell therapy and regenerative medicine strategies, stem cells have potential to treat many previously incurable diseases. However, current application methods using stem cells are underdeveloped, as these cells are used directly regardless of their culture medium and subgroup. For example, when using mesenchymal stem cells (MSCs) in cell therapy, researchers do not consider their source and culture method nor their application angle and function (soft tissue regeneration, hard tissue regeneration, suppression of immune function, or promotion of immune function). By combining machine learning methods (such as deep learning) with data sets obtained through single-cell RNA sequencing (scRNA-seq) technology, we can discover the hidden structure of these cells, predict their effects more accurately, and effectively use subpopulations with differentiation potential for stem cell therapy. scRNA-seq technology has changed the study of transcription, because it can express single-cell genes with single-cell anatomical resolution. However, this powerful technology is sensitive to biological and technical noise. The subsequent data analysis can be computationally difficult for a variety of reasons, such as denoising single cell data, reducing dimensionality, imputing missing values, and accounting for the zero-inflated nature. In this review, we discussed how deep learning methods combined with scRNA-seq data for research, how to interpret scRNA-seq data in more depth, improve the follow-up analysis of stem cells, identify potential subgroups, and promote the implementation of cell therapy and regenerative medicine measures.

Eleven grand challenges in single-cell data science

The recent boom in microfluidics and combinatorial indexing strategies, combined with low sequencing costs, has empowered single-cell sequencing technology. Thousands-or even millions-of cells analyzed in a single experiment amount to a data revolution in single-cell biology and pose unique data science problems. Here, we outline eleven challenges that will be central to bringing this emerging field of single-cell data science forward. For each challenge, we highlight motivating research questions, review prior work, and formulate open problems. This compendium is for established researchers, newcomers, and students alike, highlighting interesting and rewarding problems for the coming years.