Exploring generative deep learning for omics data using log-linear models

Hess, Moritz; Hackenberg, Maren; Binder, Harald

doi:10.1093/bioinformatics/btaa623

Cited by 8 publications

(17 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Specifically, we transfer a label of an observed cell-type to the sample, by pattern matching in the eight identified variables. As also demonstrated in Hess et al (2020) these genes carry enough information to annotate the samples generated from scVI (Fig. 8).…”

Section: Extracting Patterns With Log-linear Modelsmentioning

confidence: 56%

“…We employ the same data-set as used for the application of LDVAE. Using the approach described in Hess et al (2020), we then extract eight genes which form joint patterns with latent variables. To demonstrate that the extracted variables contribute to the essential structure in the data, we annotate samples from the posterior distribution of the VAE based on patterns identified in the eight variables.…”

Section: Extracting Patterns With Log-linear Modelsmentioning

confidence: 99%

“…Although a number of the investigated approaches allow for extracting relations between learned latent and observed variables as described e.g., in Märtens and Yau (2020), there is still a large amount of research to be performed in detecting significant associations. For instance, in both the LDVAE approach by Svensson et al (2020b), and the log-linear modeling approach by Hess et al (2020) it remains challenging to infer a cutoff defining meaningful association of observed and latent variables. Hess et al (2020) proposed to employ permutation-based approaches.…”

Section: Inferring Significance Of Statistical Associations Between Observed and Latent Variablesmentioning

confidence: 99%

“…For instance, in both the LDVAE approach by Svensson et al (2020b), and the log-linear modeling approach by Hess et al (2020) it remains challenging to infer a cutoff defining meaningful association of observed and latent variables. Hess et al (2020) proposed to employ permutation-based approaches. Even though they are effective, they are also very time-consuming and only feasible for a small number of variables.…”

Section: Inferring Significance Of Statistical Associations Between Observed and Latent Variablesmentioning

confidence: 99%

See 3 more Smart Citations

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

2022

Self Cite

View full text Add to dashboard Cite

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.

show abstract

Section: Extracting Patterns With Log-linear Modelsmentioning

confidence: 56%

Section: Extracting Patterns With Log-linear Modelsmentioning

confidence: 99%

Section: Inferring Significance Of Statistical Associations Between Observed and Latent Variablesmentioning

confidence: 99%

Section: Inferring Significance Of Statistical Associations Between Observed and Latent Variablesmentioning

confidence: 99%

See 2 more Smart Citations

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

2022

Self Cite

View full text Add to dashboard Cite

show abstract

“…We hypothesize that unsupervised generative deep learning approaches, which provide a model of the underlying data generating distribution, can be useful to build such integrated models. Specifically, we consider VAEs for this task, as they can be trained in an unsupervised way and infer a low-dimensional latent representation of the central factors of variation underlying the data, which facilitates interpretability and has been shown to be useful for capturing and extracting patterns in the data in an explainable artificial intelligence (AI) approach [21]. While VAEs have been adapted to two-photon imaging data for the specific task of inferring neural spike rates from fluorescence traces [19], we aim to exemplify how a deep learning approach based on a VAE architecture can be adapted to provide a flexible and versatile model not restricted to a specific task, yet be tailored to the properties of two-photon imaging data.…”

Section: Introductionmentioning

confidence: 99%

Incorporating structural knowledge into unsupervised deep learning for two-photon imaging data

Eichin

Hackenberg

Broichhagen

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Live imaging techniques, such as two-photon imaging, promise novel insights into cellular activity patterns at a high spatial and temporal resolution. While current deep learning approaches typically focus on specific supervised tasks in the analysis of such data, e.g., learning a segmentation mask as a basis for subsequent signal extraction steps, we investigate how unsupervised generative deep learning can be adapted to obtain interpretable models directly at the level of the video frames. Specifically, we consider variational autoencoders for models that infer a compressed representation of the data in a low-dimensional latent space, allowing for insight into what has been learned. Based on this approach, we illustrate how structural knowledge can be incorporated into the model architecture to improve model fitting and interpretability. Besides standard convolutional neural network components, we propose an architecture for separately encoding the foreground and background of live imaging data. We exemplify the proposed approach with two-photon imaging data from hippocampal CA1 neurons in mice, where we can disentangle the neural activity of interest from the neuropil background signal. Subsequently, we illustrate how to impose smoothness constraints onto the latent space for leveraging knowledge about gradual temporal changes. As a starting point for adaptation to similar live imaging applications, we provide a Jupyter notebook with code for exploration. Taken together, our results illustrate how architecture choices for deep generative models, such as for spatial structure, foreground vs. background, and gradual temporal changes, facilitate a modeling approach that combines the flexibility of deep learning with the benefits of incorporating domain knowledge. Such a strategy is seen to enable interpretable, purely image-based models of activity signals from live imaging, such as for two-photon data.

show abstract

Artificial intelligence and deep learning in molecular testing

Levy,

Vaickus

2024

Diagnostic Molecular Pathology

View full text Add to dashboard Cite

Exploring generative deep learning for omics data using log-linear models

Cited by 8 publications

References 21 publications

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

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

Incorporating structural knowledge into unsupervised deep learning for two-photon imaging data

Artificial intelligence and deep learning in molecular testing

Contact Info

Product

Resources

About