Learning to Encode Cellular Responses to Systematic Perturbations with Deep Generative Models

Xue, Yifan; Ding, Michael Q.; Lu, Xinghua

doi:10.1101/2020.01.14.906768

Cited by 5 publications

(8 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, the MMD-VAE failed to simulate all cell morphology modes, which the other VAE variants successfully captured. By training these VAE variants on L1000 gene expression readouts resulting from the same set of compound perturbations, we observed large differences in optimal hyperparameters as compared to training on Cell Painting image-based data (Subramanian et al, 2017;Xue et al, 2020). These observations indicate that the KL divergence penalty strongly influences cell morphology modeling ability, and that lessons learned by modeling other biomedical data types, such as gene expression, will not necessarily directly translate to cell morphology (Yang et al, 2021).…”

Section: Discussionmentioning

confidence: 97%

“…Recently, VAEs have been successfully trained on various biomedical data modalities such as bulk and single-cell gene expression data (Xue et al, 2020) from different assays measuring cell line perturbations and patient-derived tissue (Lotfollahi et al, 2019;Rampášek et al, 2019;Lopez et al, 2018;Way & Greene, 2018), DNA methylation (Levy et al, 2020), and cell image pixels (Lafarge et al, 2018;Ternes et al, 2021). β-VAEs have been used to produce disentangled latent representations of single cell RNA-seq data (Kimmel, 2020).…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Predicting drug polypharmacology from cell morphology readouts using variational autoencoder latent space arithmetic

Chow

Singh

Carpenter

et al. 2021

Preprint

View full text Add to dashboard Cite

A variational autoencoder (VAE) is a machine learning algorithm, useful for generating a compressed and interpretable latent space. These representations have been generated from various biomedical data types and can be used to produce realistic-looking simulated data. However, standard vanilla VAEs suffer from entangled and uninformative latent spaces, which can be mitigated using other types of VAEs such as β-VAE and MMD-VAE. In this project, we evaluated the ability of VAEs to learn cell morphology characteristics derived from cell images. We trained and evaluated these three VAE variants-Vanilla VAE, β-VAE, and MMD-VAE-on cell morphology readouts and explored the generative capacity of each model to predict compound polypharmacology (the interactions of a drug with more than one target) using an approach called latent space arithmetic (LSA). To test the generalizability of the strategy, we also trained these VAEs using gene expression data of the same compound perturbations and found that gene expression provides complementary information. We found that the β-VAE and MMD-VAE disentangle morphology signals and reveal a more interpretable latent space. We reliably simulated morphology and gene expression readouts from certain compounds thereby predicting cell states perturbed with compounds of known polypharmacology. Inferring cell state for specific drug mechanisms could aid researchers in developing and identifying targeted therapeutics and categorizing off-target effects in the future.

show abstract

Section: Discussionmentioning

confidence: 97%

Section: Introductionmentioning

confidence: 99%

Predicting drug polypharmacology from cell morphology readouts using variational autoencoder latent space arithmetic

Chow

Singh

Carpenter

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…To obtain an integrative drug embedding (IDE) reflecting multiple aspects of the drug, we adopted a semi-supervised method 19 that integrated two sources of information: 1) chemical structure of the drug, and 2) functional impact of the drug on gene expression. 17 To represent chemical structures, we used the SMILES 22 representation of molecules and obtained SMILES strings of 250K drug-like molecules from the ZINC 23 database. To represent the functional impact on gene expression, we trained a variational auto-encoder (VAE) 24 model on the transcriptomic data of cell lines treated by different drugs from the LINCS database.…”

Section: Methodsmentioning

confidence: 99%

“…Prior approaches to learning representations of drugs have transformed their chemical structures from a form defined by the simplified molecularinput line-entry system (SMILES) into a vector (an embedding) that can be concatenated with cell embeddings in a deep learning model to predict drug response. [14][15][16] However, this approach does not utilize a rich body of information regarding the functional impact of chemicals on cell signaling systems, 8,9,17 which is highly relevant to the MOAs of drugs 18 and thus relevant to predicting drug responses.…”

Section: Introductionmentioning

confidence: 99%

“…The main innovation of our model lies in integration of multiple sources of information and different learning techniques, including: 1) Integrative representation of drugs . For the representation of drugs, we combined three kinds of information regarding a drug — its chemical structure, 19 its impact on cellular transcriptomic signaling, 17 and its pathway information 20 — to make the representation more informative. 2) Representation learning through collaborative filtering .…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

De novo Prediction of Cell-Drug Sensitivities Using Deep Learning-based Graph Regularized Matrix Factorization

Ren

Tao²,

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Application of artificial intelligence (AI) in precision oncology typically involves predicting whether the cancer cells of a patient (previously unseen by AI models) will respond to any of a set of existing anticancer drugs, based on responses of previous training cell samples to those drugs. To expand the repertoire of anticancer drugs, AI has also been used to repurpose drugs that have not been tested in an anticancer setting, i.e., predicting the anticancer effects of a new drug on previously unseen cancer cells de novo. Here, we report a computational model that addresses both of the above tasks in a unified AI framework. Our model, referred to as deep learning-based graph regularized matrix factorization (DeepGRMF), integrates neural networks, graph models, and matrix-factorization techniques to utilize diverse information from drug chemical structures, their impact on cellular signaling systems, and cancer cell cellular states to predict cell response to drugs. DeepGRMF learns embeddings of drugs so that drugs sharing similar structures and mechanisms of action (MOAs) are closely related in the embedding space. Similarly, DeepGRMF also learns representation embeddings of cells such that cells sharing similar cellular states and drug responses are closely related. Evaluation of DeepGRMF and competing models on Genomics of Drug Sensitivity in Cancer (GDSC) and Cancer Cell Line Encyclopedia (CCLE) datasets show its superiority in prediction performance. Finally, we show that the model is capable of predicting effectiveness of a chemotherapy regimen on patient outcomes for the lung cancer patients in The Cancer Genome Atlas (TCGA) dataset.

show abstract

Evolution of precision oncology‐guided treatment paradigms

DasGupta

Yap

Yaqing

et al. 2022

WIREs Mechanisms of Disease

View full text Add to dashboard Cite

Cancer treatment is gradually evolving from the classical use of nonspecific cytotoxic drugs targeting generic mechanisms of cell growth and proliferation. Instead, new “patient‐specific treatment paradigms” that are based on an individual patient's tumor‐specific molecular features are emerging, and these include “druggable” genomic alterations such as oncogenic driver mutations, downstream activities of cancer‐signaling pathways, and the expression of specific genes involved in tumorigenesis and cancer progression. This evolving landscape of making evidence‐based treatment decisions forms the foundation of precision oncology, which aims to deliver “the right drug, to the right patient and at the right time”. The long‐term vision for this approach is to maximize the treatment efficacy while minimizing exposure to ineffective therapy and reducing co‐morbidity‐related side effects. Successful clinical translation and implementation of this vision have the potential to revolutionize treatment paradigms from predominantly reactive, to more evidence‐based, proactive and predictive care. In this article, we review the past and current approaches in precision oncology, and describe their remarkable power and limitations. We also speculate on the evolution of newly emerging methodologies of the future that can be used to address some of the key challenges associated with the existing paradigms. This article is categorized under: Cancer > Genetics/Genomics/Epigenetics Cancer > Molecular and Cellular Physiology Cancer > Computational Models

show abstract

Learning to Encode Cellular Responses to Systematic Perturbations with Deep Generative Models

Cited by 5 publications

References 29 publications

Predicting drug polypharmacology from cell morphology readouts using variational autoencoder latent space arithmetic

Predicting drug polypharmacology from cell morphology readouts using variational autoencoder latent space arithmetic

De novo Prediction of Cell-Drug Sensitivities Using Deep Learning-based Graph Regularized Matrix Factorization

Evolution of precision oncology‐guided treatment paradigms

Contact Info

Product

Resources

About