Generative modeling of single-cell gene expression for dose-dependent chemical perturbations

Kana, Omar; Nault, Rance; Filipovic, David; Marri, Daniel; Zacharewski, Tim; Bhattacharya, Sudin

doi:10.1016/j.patter.2023.100817

Cited by 11 publications

(9 citation statements)

References 53 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To showcase the capability of EXPORT in unraveling the ordinal intricacies of perturbed transcriptomics data on single-cell resolution, we select the single nuclei RNA sequencing (snRNA-seq) data of mouse liver cells gavaged with 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) at doses of 0.01, 0.03, 0.1, 0.3, 1, 3, 10, or 30 μg/kg (Kana et al, 2023; Nault et al, 2022). This dataset is publicly available in the Gene Expression Omnibus (GEO) under the accession number GSE184506.…”

Section: Dataset Detailsmentioning

confidence: 99%

“…A total of 131,613 nuclei are sequenced in this dataset. As due to the uneven expression of the TCDD canonical receptor, the aryl hydrocarbon receptor (AhR), the hepatic responses to TCDD vary across different cell types in the liver and also within cell types (such as hepatocytes), in our analysis with EXPORT, we only consider cells from hepatocyte-portals (Kana et al, 2023). This leads to selection of 57,284 nuclei expression profiles from hepatocyte-portal cell types.…”

Section: Dataset Detailsmentioning

confidence: 99%

“…Of particular interest are the changes of expression patterns as responses to ordinal perturbations, such as drug screening and radiation exposure at different dosage levels (Peidli et al, 2024; Luo et al, 2022). The cellular responses to external stimuli are intricately linked to the ordinal dosage of the stimulus, delineate dose-response curves that provide invaluable insights into the dynamics of biological systems (Kana et al, 2023). Exploring and understanding how cellular responses evolve across a spectrum of stimulus doses sheds light on the underlying cellular mechanisms and paves the way for precision medicine approaches tailored to individualized dosedependent responses specifically in the context of disease and drug treatment (Bock et al, 2022; Lotfollahi et al, 2019).…”

Section: Introductionmentioning

confidence: 99%

“…Deep generative models such as VAEs (Kingma & Welling, 2014) have shown their capabilities in unraveling biological insights from large and heterogeneous perturbation-induced gene expression profiles due to their model flexibility while keeping necessary changes to the inference procedure relatively minimal (Kana et al, 2023; Lotfollahi et al, 2023). Despite these successes, VAEs suffer from limited interpretability, making them ‘black boxes’ as the reasoning behind predictions as well as the direct correspondence between input features, learned latent vectors and biological processes is often unknown (Lopez et al, 2023).…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Biologically Interpretable VAE with Supervision for Transcriptomics Data Under Ordinal Perturbations

Niyakan,

Yoon,

Qian

et al. 2024

Preprint

View full text Add to dashboard Cite

Latent variable models such as the Variational Auto-Encoders (VAEs) have shown impressive performance for inferring expression patterns for cell subtyping and biomarker identification from transcriptomics data. However, the limited interpretability of their latent variables obscures deriving meaningful biological understanding of cellular responses to different external and internal perturbations. We here propose a novel deep learning framework, EXPORT (EXPlainable VAE for ORdinally perturbed Transcriptomics data), for analyzing ordinally perturbed transcriptomics data that can incorporate any biological pathway knowledge in the VAE latent space. With the corresponding pathway-informed decoder, the learned latent expression patterns can be explained as pathway-level responses to perturbations, offering direct interpretability with biological understanding. More importantly, we explicitly model the ordinal nature of many real-world perturbations into the EXPORT framework by training an auxiliary ordinal regressor neural network to capture corresponding expression changes in the VAE latent representations, for example under different dosage levels of radiation exposure. By incorporating ordinal constraints during the training of our proposed framework, we further enhance the model interpretability by guiding the VAE latent space to organize perturbation responses in a hierarchical manner. We demonstrate the utility of the inferred guided latent space for downstream tasks, such as identifying key regulatory pathways associated with specific perturbation changes by analyzing transcriptomics datasets on both bulk and single-cell data. Overall, we envision that our proposed approach can unravel unprecedented biological intricacies in cellular responses to various perturbations while bringing an additional layer of interpretability to biology-inspired deep learning models.

show abstract

Section: Dataset Detailsmentioning

confidence: 99%

Section: Dataset Detailsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Biologically Interpretable VAE with Supervision for Transcriptomics Data Under Ordinal Perturbations

Niyakan,

Yoon,

Qian

et al. 2024

Preprint

View full text Add to dashboard Cite

show abstract

“…scGen has the capability to deduce data points for conditions not present in the data. It successfully predicted LPS responses across various animal species 27 , as well as protein abundance from mRNA in cases of Alzheimer’s dementia 31 , and gene expression patterns in mice, humans, and cancer cell-lines following chemical perturbations 32 .…”

Section: Introductionmentioning

confidence: 99%

Neural Network-Assisted Humanization of COVID-19 Hamster scRNAseq Data Reveals Matching Severity States in Human Disease

Friedrich,

Pennitz,

Wyler

et al. 2024

Preprint

View full text Add to dashboard Cite

Translating findings from animal models to human disease is essential for dissecting disease mechanisms, developing and testing precise therapeutic strategies. The coronavirus disease 2019 (COVID-19) pandemic has highlighted this need, particularly for models showing disease severity-dependent immune responses. Single-cell transcriptomics (scRNAseq) is well poised to reveal similarities and differences between species at the molecular and cellular level with unprecedented resolution. However, computational methods enabling detailed matching are still scarce. Here, we provide a structured scRNAseq-based approach that we applied to scRNAseq from blood leukocytes originating from humans and hamsters affected with moderate or severe COVID-19. Integration of COVID-19 patient data with two hamster models that develop moderate (Syrian hamster, Mesocricetus auratus) or severe (Roborovski hamster, Phodopus roborovskii) disease revealed that most cellular states are shared across species. A neural network-based analysis using variational autoencoders quantified the overall transcriptomic similarity across species and severity levels, showing highest similarity between neutrophils of Roborovski hamsters and severe COVID-19 patients, while Syrian hamsters better matched patients with moderate disease, particularly in classical monocytes. We further used transcriptome-wide differential expression analysis to identify which disease stages and cell types display strongest transcriptional changes. Consistently, hamster's response to COVID-19 was most similar to humans in monocytes and neutrophils. Disease-linked pathways found in all species specifically related to interferon response or inhibition of viral replication. Analysis of candidate genes and signatures supported the results. Our structured neural network-supported workflow could be applied to other diseases, allowing better identification of suitable animal models with similar pathomechanisms across species.

show abstract

Artificial intelligence and deep learning in single-cell omics data analysis: A case study

Anjum,

Khalid,

Takkar

et al. 2025

Deep Learning in Genetics and Genomics

View full text Add to dashboard Cite

Generative modeling of single-cell gene expression for dose-dependent chemical perturbations

Cited by 11 publications

References 53 publications

Biologically Interpretable VAE with Supervision for Transcriptomics Data Under Ordinal Perturbations

Biologically Interpretable VAE with Supervision for Transcriptomics Data Under Ordinal Perturbations

Neural Network-Assisted Humanization of COVID-19 Hamster scRNAseq Data Reveals Matching Severity States in Human Disease

Artificial intelligence and deep learning in single-cell omics data analysis: A case study

Contact Info

Product

Resources

About