Daniel Phillip Lewinsohn
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Daniel Phillip Lewinsohn has a total of 22 co-authors. Here are the top co-authors by number of shared publications:
- Allon Wagner
Center for Computational Biology, University of California, Berkeley, Berkeley, CA, USA
2 shared publications - Donald F. Conrad
Oregon Health and Science University
2 shared publications Katinka A Vigh-Conrad
Division of Genetics, Oregon National Primate Research Center, Oregon Health and Science University , Beaverton, OR 97006, United States
2 shared publications- Nicolas Dias
Department of Genetics and the Yale Stem Cell Center, Yale School of Medicine, Yale University, New Haven, CT, USA
2 shared publications Adelina Chau
Department of Electrical Engineering and Computer Sciences, University of California, Berkeley, Berkeley, CA, USA
1 shared publication- Adelina Chau
Department of Electrical Engineering and Computer Sciences, University of California, Berkeley, Berkeley, CA, USA
1 shared publication
Latest Publications
- Open Access
Extracting interpretable single-cell metabolic states with graph-guided representation learning
Summary: Metabolism shapes cellular function and state, yet measuring single-cell metabolic states at scale remains a challenge. We present Metabolic Representation Net (MeRN), a graph-guided variational autoencoder that leverages prior metabolic knowledge as a topology graph to learn latent representations of metabolic state and reaction activity from single-cell transcriptomes. MeRN’s scalable estimation of reaction activity enables the definition of data-driven pathways (DDPs): context-specific metabolic modules supported by transcriptomic evidence and agnostic of standard pathway definitions. Using DDPs, we introduce the weakest link analysis to identify metabolic network rewiring. MeRN recovers metabolic zonation in the mouse intestine, links a folate deficiency-induced break in de novo purine synthesis to embryonic neural tube defects, shows cytokines with similar non-metabolic effects can elicit divergent T cell metabolism, and identifies metabolic drivers of T cell exhaustion and therapy response in human cancers. Our results establish MeRN as a unified method for metabolic analysis of single-cell transcriptomes. Research highlights: MeRN leverages the metabolic topology to comprehensively predict reaction- and pathway-level metabolic activities MeRN enables data-driven pathways (DDPs) that capture empirically supported cell-type-specific metabolic modules, agnostic of standard pathway definitions MeRN-based DDPs identify metabolic network rewiring of de novo purine synthesis due to folate deficiency during embryonic neural tube development MeRN identifies human pan-cancer metabolic drivers of T cell exhaustion and Treg-specific metabolic adaptations
Open AccessExtracting interpretable single-cell metabolic states with graph-guided representation learning
Summary: Metabolism shapes cellular function and state, yet measuring single-cell metabolic states at scale remains a challenge. We present Metabolic Representation Net (MeRN), a graph-guided variational autoencoder that leverages prior metabolic knowledge as a topology graph to learn latent representations of metabolic state and reaction activity from single-cell transcriptomes. MeRN’s scalable estimation of reaction activity enables the definition of data-driven pathways (DDPs): context-specific metabolic modules supported by transcriptomic evidence and agnostic of standard pathway definitions. Using DDPs, we introduce the weakest link analysis to identify metabolic network rewiring. MeRN recovers metabolic zonation in the mouse intestine, links a folate deficiency-induced break in de novo purine synthesis to embryonic neural tube defects, shows cytokines with similar non-metabolic effects can elicit divergent T cell metabolism, and identifies metabolic drivers of T cell exhaustion and therapy response in human cancers. Our results establish MeRN as a unified method for metabolic analysis of single-cell transcriptomes. Research highlights: MeRN leverages the metabolic topology to comprehensively predict reaction- and pathway-level metabolic activities MeRN enables data-driven pathways (DDPs) that capture empirically supported cell-type-specific metabolic modules, agnostic of standard pathway definitions MeRN-based DDPs identify metabolic network rewiring of de novo purine synthesis due to folate deficiency during embryonic neural tube development MeRN identifies human pan-cancer metabolic drivers of T cell exhaustion and Treg-specific metabolic adaptations
- Open Access
Folate deficiency disrupts key metabolic transitions within the developing neural ectoderm
Summary: Despite long-standing epidemiological associations, the mechanism linking folate availability to gestational neural tube defects remains unclear, partly because measuring and interpreting metabolic activity in dynamic biological systems remains challenging. Here, we apply a deep-learning-based graph-guided variational autoencoder (MeRN; Metabolic Representation Network) to infer single-cell metabolic activity and states from scRNA-seq data of mouse embryogenesis. By analyzing folate-deficient embryogenesis from E7.0–E9.0, we identify a transient state within the nascent neural lineage that is acutely sensitive to folate availability, leading to an interconnected disruption between key bioenergetic pathways and de novo purine biosynthesis. Moreover, metabolically induced growth defects lead to permanent morphological disruptions along the dorsal-ventral axis, which we confirm by generating whole-embryo fate maps using a prime-editing-based lineage recorder (PEtracer). Collectively, our results establish a highly scalable framework for interpreting dynamic changes in embryonic metabolism and elucidating the mechanistic bases underlying environmentally linked congenital disorders. Research highlights: Developing embryos adopt metabolic states that are distinct from cell lineage Metabolic states are organized temporally and spatially along key body axes Neural tube development depends on a rapid transition between metabolic states Activation of transient neural metabolic states is highly sensitive to disrupted folate absorption The floor plate is preferentially resistant to folate deprivation and distorts subsequent neural patterning
Open AccessFolate deficiency disrupts key metabolic transitions within the developing neural ectoderm
Summary: Despite long-standing epidemiological associations, the mechanism linking folate availability to gestational neural tube defects remains unclear, partly because measuring and interpreting metabolic activity in dynamic biological systems remains challenging. Here, we apply a deep-learning-based graph-guided variational autoencoder (MeRN; Metabolic Representation Network) to infer single-cell metabolic activity and states from scRNA-seq data of mouse embryogenesis. By analyzing folate-deficient embryogenesis from E7.0–E9.0, we identify a transient state within the nascent neural lineage that is acutely sensitive to folate availability, leading to an interconnected disruption between key bioenergetic pathways and de novo purine biosynthesis. Moreover, metabolically induced growth defects lead to permanent morphological disruptions along the dorsal-ventral axis, which we confirm by generating whole-embryo fate maps using a prime-editing-based lineage recorder (PEtracer). Collectively, our results establish a highly scalable framework for interpreting dynamic changes in embryonic metabolism and elucidating the mechanistic bases underlying environmentally linked congenital disorders. Research highlights: Developing embryos adopt metabolic states that are distinct from cell lineage Metabolic states are organized temporally and spatially along key body axes Neural tube development depends on a rapid transition between metabolic states Activation of transient neural metabolic states is highly sensitive to disrupted folate absorption The floor plate is preferentially resistant to folate deprivation and distorts subsequent neural patterning
- Open Access
- 1 June 2023
- Published by Oxford University Press (OUP) in Bioinformatics
- Vol. 39 (6)
Consensus label propagation with graph convolutional networks for single-cell RNA sequencing cell type annotation
Motivation: Single-cell RNA sequencing (scRNA-seq) data, annotated by cell type, is useful in a variety of downstream biological applications, such as profiling gene expression at the single-cell level. However, manually assigning these annotations with known marker genes is both time-consuming and subjective. Results: We present a Graph Convolutional Network (GCN)-based approach to automate the annotation process. Our process builds upon existing labeling approaches, using state-of-the-art tools to find cells with highly confident label assignments through consensus and spreading these confident labels with a semi-supervised GCN. Using simulated data and two scRNA-seq datasets from different tissues, we show that our method improves accuracy over a simple consensus algorithm and the average of the underlying tools. We also compare our method to a nonparametric neighbor majority approach, showing comparable results. We then demonstrate that our GCN method allows for feature interpretation, identifying important genes for cell type classification. We present our completed pipeline, written in PyTorch, as an end-to-end tool for automating and interpreting the classification of scRNA-seq data. Availability and implementation: Our code for conducting the experiments in this paper and using our model is available at https://github.com/lewinsohndp/scSHARP.
- 1 June 2023
- Published by Oxford University Press (OUP) in Bioinformatics
- Vol. 39 (6)
Open AccessConsensus label propagation with graph convolutional networks for single-cell RNA sequencing cell type annotation
Motivation: Single-cell RNA sequencing (scRNA-seq) data, annotated by cell type, is useful in a variety of downstream biological applications, such as profiling gene expression at the single-cell level. However, manually assigning these annotations with known marker genes is both time-consuming and subjective. Results: We present a Graph Convolutional Network (GCN)-based approach to automate the annotation process. Our process builds upon existing labeling approaches, using state-of-the-art tools to find cells with highly confident label assignments through consensus and spreading these confident labels with a semi-supervised GCN. Using simulated data and two scRNA-seq datasets from different tissues, we show that our method improves accuracy over a simple consensus algorithm and the average of the underlying tools. We also compare our method to a nonparametric neighbor majority approach, showing comparable results. We then demonstrate that our GCN method allows for feature interpretation, identifying important genes for cell type classification. We present our completed pipeline, written in PyTorch, as an end-to-end tool for automating and interpreting the classification of scRNA-seq data. Availability and implementation: Our code for conducting the experiments in this paper and using our model is available at https://github.com/lewinsohndp/scSHARP.
- Open Access
Consensus Label Propagation with Graph Convolutional Networks for Single-Cell RNA Sequencing Cell Type Annotation
Motivation: Single-cell RNA sequencing (scRNA-seq) data, annotated by cell type, is useful in a variety of downstream biological applications, such as profiling gene expression at the single-cell level. However, manually assigning these annotations with known marker genes is both time-consuming and subjective. Results: We present a Graph Convolutional Network (GCN) based approach to automate the annotation process. Our process builds upon existing labeling approaches, using state-of-the-art tools to find cells with highly confident label assignments through consensus and spreading these confident labels with a semi-supervised GCN. Using simulated data and two scRNA-seq data sets from different tissues, we show that our method improves accuracy over a simple consensus algorithm and the average of the underlying tools. We also compare our method to a non-parametric neighbor majority approach, showing comparable results. We then demonstrate that our GCN method allows for feature interpretation, identifying important genes for cell type classification. We present our completed pipeline, written in PyTorch, as an end-to-end tool for automating and interpreting the classification of scRNA-seq data. Availability: Our code for conducting the experiments in this paper and using our model is available at https://github.com/lewinsohndp/scSHARP Contact: d_lewinsohn@coloradocollege.edu Supplementary information: Supplementary data are available at Journal Name online.
Open AccessConsensus Label Propagation with Graph Convolutional Networks for Single-Cell RNA Sequencing Cell Type Annotation
Motivation: Single-cell RNA sequencing (scRNA-seq) data, annotated by cell type, is useful in a variety of downstream biological applications, such as profiling gene expression at the single-cell level. However, manually assigning these annotations with known marker genes is both time-consuming and subjective. Results: We present a Graph Convolutional Network (GCN) based approach to automate the annotation process. Our process builds upon existing labeling approaches, using state-of-the-art tools to find cells with highly confident label assignments through consensus and spreading these confident labels with a semi-supervised GCN. Using simulated data and two scRNA-seq data sets from different tissues, we show that our method improves accuracy over a simple consensus algorithm and the average of the underlying tools. We also compare our method to a non-parametric neighbor majority approach, showing comparable results. We then demonstrate that our GCN method allows for feature interpretation, identifying important genes for cell type classification. We present our completed pipeline, written in PyTorch, as an end-to-end tool for automating and interpreting the classification of scRNA-seq data. Availability: Our code for conducting the experiments in this paper and using our model is available at https://github.com/lewinsohndp/scSHARP Contact: d_lewinsohn@coloradocollege.edu Supplementary information: Supplementary data are available at Journal Name online.