Single-Cell Techniques and Deep Learning in Predicting Drug Response

Wu, Zhenyu; Lawrence, Patrick J.; Ma, Anjun; Zhu, Jian; Dong, Xu; Ma, Qin

doi:10.1016/j.tips.2020.10.004

Cited by 38 publications

(24 citation statements)

References 84 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the same way, with the many and large single-cell data now being generated (Svensson et al, 2020), which simultaneously profile hundreds of perturbations, it becomes possible for DL models to identify patterns of perturbation effect. Indeed, the flexibility of deep neural networks has often allowed them to outperform classical ML in situations with increasing data size and complexity, also in biology (Wu et al, 2020).…”

Section: Of Patterns In Perturbation Biologymentioning

confidence: 99%

“…This manuscript focuses on more recent developments that have emerged as the increasing availability of high-throughput multi-omic data has made it possible to leverage DL methods to establish more fine-grained and predictive models for the above tasks (Zhou and Troyanskaya, 2015; Eraslan et al, ll 2019; Zheng and Wang, 2019;Wu et al, 2020). In particular, single-cell profiling methods that generate data with a high number of observations enable training DL models on the transcriptional, proteomic, and epigenetic level, which was previously untenable due to the low number of observations (Hua et al, 2005).…”

Section: Introductionmentioning

confidence: 99%

“…This high sample size makes it possible to simultaneously model heterogeneous cellular responses along multiple axes of variation. The foremost example is single-cell RNA sequencing (scRNA-seq), which captures the full heterogeneity of response in hundreds of thousands of cells, augmenting previously low-dimensional data like drug-cell line viability and dose response curves (Goldman et al, 2019;Wu et al, 2020). Because they provide more examples and more variation, single-cell data potentially allow models to extrapolate to unseen events, with fewer experiments.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Machine learning for perturbational single-cell omics

Lotfollahi

Wolf

et al. 2021

Cell Systems

View full text Add to dashboard Cite

Section: Of Patterns In Perturbation Biologymentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Machine learning for perturbational single-cell omics

Lotfollahi

Wolf

et al. 2021

Cell Systems

View full text Add to dashboard Cite

“…As far as we know, only six datasets with drug treatment and experimentally validated drug responses for individual cells are available in the public domain. Fortunately, bulk drug-related RNA-Seq data can be great complementary resources to infer relations of gene expression-drug response to help predict drug responses at the single-cell level (7), if bulk and single-cell data can be integrated (8).…”

Section: Introductionmentioning

confidence: 99%

Deep Transfer Learning of Drug Responses by Integrating Bulk and Single-cell RNA-seq data

Chen

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Massively bulk RNA sequencing databases incorporating drug screening have opened up an avenue to inform the optimal clinical application of cancer drugs. Meanwhile, the growing single-cell RNA sequencing data contributes to improving therapeutic effectiveness by studying the heterogeneity of drug responses for cancer cell subpopulations. Yet, the drug response information for single-cell data is scarcely obtained. Thus, there is an urgent need to develop computational pipelines to infer and interpret cancer drug responses in single cells. Here, we developed scDEAL, a deep transfer learning framework integrating large-scale bulk and single-cell RNA sequencing drug response datasets. We benchmarked scDEAL on six single-cell RNA sequencing datasets and indicate its model interpretability by several case studies. scDEAL not only achieves accurate and robust performance in single-cell drug response predictions, but also can infer signature genes to reveal potential drug resistance mechanisms based on integrated gradient feature interpretation. This work may help study cell reprogramming, drug selection, and repurposing for improving therapeutic efficacy.

show abstract

“…Consequently, fully considering both reversed expressions and adverse effects is a highly desirable strategy [9,10]. What is more, as cancer cells may develop drug resistance, it has been suggested that treatments that employ drug combinations potentially enhance efficacy and reduce toxicity [11]. Therefore, the mechanism of synergy and new combination recommendations have acted as a catalyst for intensive studies by academic researchers and pharmaceutical enterprises.…”

Section: Introductionmentioning

confidence: 99%

DBPOM: a comprehensive database of pharmaco-omics for cancer precision medicine

Liu

Song

et al. 2021

Preprint

View full text Add to dashboard Cite

During the course of cancer treatment, both efficacy and the adverse effects of drugs on patients should be taken into account. Although some public databases and modeling frameworks have been developed through studies on drug response, the negative effects of drugs are always neglected. Furthermore, most of them only considered the ramifications of the drug on the cell line, but the effects on the patient still require a huge amount of work to integrate data from various databases and calculations, especially in relation to precision treatment. In order to address these issues, we developed the DBPOM (http://www.dbpom.net/, a comprehensive database of pharmaco-omics for cancer precision medicine), which explores various drugs’ efficacy levels by calculating their potency in reverse, or enhancing cancer-associated gene expression changes. When compared with existing databases, the DBPOM could estimate the effectiveness of a drug on individual patients through the mapping of various cell lines to each person according to their genetic mutation similarities. The DBPOM is an easy-to-use and one-stop database for clinicians and drug researchers to search and analyze the overall effect of a drug or a drug combination on cancer patients as well as the biological functions that they target. We anticipate that DBPOM will become an important resource and analysis platform for drug development, drug mechanism studies and the discovery of new therapies.

show abstract

Single-Cell Techniques and Deep Learning in Predicting Drug Response

Cited by 38 publications

References 84 publications

Machine learning for perturbational single-cell omics

Machine learning for perturbational single-cell omics

Deep Transfer Learning of Drug Responses by Integrating Bulk and Single-cell RNA-seq data

DBPOM: a comprehensive database of pharmaco-omics for cancer precision medicine

Contact Info

Product

Resources

About