Sehi L’ Yi
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
Sehi L’ Yi has a total of 273 co-authors. Here are the top co-authors by number of shared publications:
- Nils Gehlenborg
Department of Biomedical Informatics, Harvard Medical School, Boston, MA, USA
27 shared publications - Isaac S. Kohane
Department of Biomedical Informatics, Harvard Medical School, Boston, Massachusetts, USA
7 shared publications - Tianxi Cai
Department of Biomedical Informatics, Harvard Medical School, Boston, Massachusetts, USA
7 shared publications - Alba Gutiérrez-Sacristán
Department of Biomedical Informatics, Harvard Medical School, Boston, MA, USA
6 shared publications - David A. Hanauer
Michigan Institute for Clinical and Health Research, University of Michigan Medical School , Ann Arbor, MI, 48109
6 shared publications - Gabriel. A. Brat
Department of Surgery, Beth Israel Deaconess Medical Center, Boston, MA 02215, USA
6 shared publications
Latest Publications
- 1 July 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Visualization and Computer Graphics
- Vol. 32 (7) , 6447-6463
Geranium: Multimodal Retrieval of Genomics Data Visualizations
Effective visualization is essential for interpreting genomics data, yet researchers often face challenges in finding relevant, reusable examples. Existing tools offer limited support for searching the vast landscape of genomics visualizations, making the process of authoring new visualizations time-consuming and inefficient. To address this gap, we introduce Geranium, a data visualization retrieval system for searching and authoring genomics visualizations. Geranium supports multimodal retrieval, enabling users to query with images, text, or grammar-based specifications. Retrieved examples serve as scaffolds for authoring, providing templates that researchers can adapt with their own data, thereby streamlining the mechanics of visualization construction. Geranium integrates three embedding methods to combine specialized and general knowledge: grammar-based embeddings tailored to genomics visualizations, multimodal embeddings from a biomedical vision-language foundation model, and text embeddings from a fine-tuned large language model. For each visualization, we construct a multimodal representation that includes a Gosling specification, a pixel-based rendering, and natural language descriptions. We evaluate embedding strategies to maximize top-$k$k retrieval accuracy and conduct user studies with domain collaborators to gather feedback on usability. Our collection comprises 3,200 visualizations across 50 categories, ranging from single-view to coordinated multi-view designs and supporting applications from single-cell epigenomics to structural variation analysis.
- 1 July 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Visualization and Computer Graphics
- Vol. 32 (7) , 6447-6463
Geranium: Multimodal Retrieval of Genomics Data Visualizations
Effective visualization is essential for interpreting genomics data, yet researchers often face challenges in finding relevant, reusable examples. Existing tools offer limited support for searching the vast landscape of genomics visualizations, making the process of authoring new visualizations time-consuming and inefficient. To address this gap, we introduce Geranium, a data visualization retrieval system for searching and authoring genomics visualizations. Geranium supports multimodal retrieval, enabling users to query with images, text, or grammar-based specifications. Retrieved examples serve as scaffolds for authoring, providing templates that researchers can adapt with their own data, thereby streamlining the mechanics of visualization construction. Geranium integrates three embedding methods to combine specialized and general knowledge: grammar-based embeddings tailored to genomics visualizations, multimodal embeddings from a biomedical vision-language foundation model, and text embeddings from a fine-tuned large language model. For each visualization, we construct a multimodal representation that includes a Gosling specification, a pixel-based rendering, and natural language descriptions. We evaluate embedding strategies to maximize top-$k$k retrieval accuracy and conduct user studies with domain collaborators to gather feedback on usability. Our collection comprises 3,200 visualizations across 50 categories, ranging from single-view to coordinated multi-view designs and supporting applications from single-cell epigenomics to structural variation analysis.
- Open Access
- 13 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 1-17
Automatic Synthesis of Visualization Design Knowledge Bases
- 13 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 1-17
Open AccessAutomatic Synthesis of Visualization Design Knowledge Bases
- Open Access
- 10 April 2026
- Published by Center for Open Science in OSF Preprints
Geranium: Multimodal Retrieval of Genomics Data Visualizations
Effective visualization is essential for interpreting genomics data, yet researchers often face challenges in finding relevant, reusable examples. Existing tools offer limited support for searching the vast landscape of genomics visualizations, making the process of authoring new visualizations time-consuming and inefficient. To address this gap, we introduce Geranium, a data visualization retrieval system for searching and authoring genomics visualizations. Geranium supports multimodal retrieval, enabling users to query with images, text, or grammar-based specifications. Retrieved examples serve as scaffolds for authoring, providing templates that researchers can adapt with their own data, thereby streamlining the mechanics of visualization construction. Geranium integrates three embedding methods to combine specialized and general knowledge: grammar-based embeddings tailored to genomics visualizations, multimodal embeddings from a biomedical vision-language foundation model, and text embeddings from a fine-tuned large language model. For each visualization, we construct a multimodal representation that includes a Gosling specification, a pixel-based rendering, and natural language descriptions. We evaluate embedding strategies to maximize top-k retrieval accuracy and conduct user studies with domain collaborators to gather feedback on usability. Our collection comprises 3,200 visualizations across 50 categories, ranging from single-view to coordinated multi-view designs and supporting applications from single-cell epigenomics to structural variation analysis.
- 10 April 2026
- Published by Center for Open Science in OSF Preprints
Open AccessGeranium: Multimodal Retrieval of Genomics Data Visualizations
Effective visualization is essential for interpreting genomics data, yet researchers often face challenges in finding relevant, reusable examples. Existing tools offer limited support for searching the vast landscape of genomics visualizations, making the process of authoring new visualizations time-consuming and inefficient. To address this gap, we introduce Geranium, a data visualization retrieval system for searching and authoring genomics visualizations. Geranium supports multimodal retrieval, enabling users to query with images, text, or grammar-based specifications. Retrieved examples serve as scaffolds for authoring, providing templates that researchers can adapt with their own data, thereby streamlining the mechanics of visualization construction. Geranium integrates three embedding methods to combine specialized and general knowledge: grammar-based embeddings tailored to genomics visualizations, multimodal embeddings from a biomedical vision-language foundation model, and text embeddings from a fine-tuned large language model. For each visualization, we construct a multimodal representation that includes a Gosling specification, a pixel-based rendering, and natural language descriptions. We evaluate embedding strategies to maximize top-k retrieval accuracy and conduct user studies with domain collaborators to gather feedback on usability. Our collection comprises 3,200 visualizations across 50 categories, ranging from single-view to coordinated multi-view designs and supporting applications from single-cell epigenomics to structural variation analysis.
- 15 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Visualization and Computer Graphics
- Vol. 32 (1) , 890-900
Design Space and Declarative Grammar for 3D Genomic Data Visualization
Various computational approaches predict chromatin structure, yielding concrete models that position genomic loci in physical space and help reveal genome organization and function. While prior visualization research has explored data and task abstractions for genomics, the design space for depicting these three-dimensional (3D) genome models—and associated genome mapped data—remains unclear. In this paper, we investigate the visualization of genomic data with a spatial component. First, we systematically survey how 3D genome models are used and depicted in computational biology. We analyze over 300 papers with figures that visualize 3D genomic data and categorize the methods for visual representation. From this survey, we derive a design space for visualizing 3D genome data, identifying common patterns and key properties such as representation, visual channels, and composition. We position these findings within an existing genomics visualization taxonomy, refining and extending existing classifications. Second, we augment Gosling, a declarative visualization grammar for genomics, to support 3D genomic data. Our integration enables expressive authoring of visualizations that connect traditional genome-mapped information with 3D genome models, emphasizing their spatial characteristics. To demonstrate its utility, we employ our extended grammar to recreate interactive examples, showcasing its ability to represent complex visual designs. Comprehensive examples and an interactive editor are available at 3d.gosling-lang.org.
- 15 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Visualization and Computer Graphics
- Vol. 32 (1) , 890-900
Design Space and Declarative Grammar for 3D Genomic Data Visualization
Various computational approaches predict chromatin structure, yielding concrete models that position genomic loci in physical space and help reveal genome organization and function. While prior visualization research has explored data and task abstractions for genomics, the design space for depicting these three-dimensional (3D) genome models—and associated genome mapped data—remains unclear. In this paper, we investigate the visualization of genomic data with a spatial component. First, we systematically survey how 3D genome models are used and depicted in computational biology. We analyze over 300 papers with figures that visualize 3D genomic data and categorize the methods for visual representation. From this survey, we derive a design space for visualizing 3D genome data, identifying common patterns and key properties such as representation, visual channels, and composition. We position these findings within an existing genomics visualization taxonomy, refining and extending existing classifications. Second, we augment Gosling, a declarative visualization grammar for genomics, to support 3D genomic data. Our integration enables expressive authoring of visualizations that connect traditional genome-mapped information with 3D genome models, emphasizing their spatial characteristics. To demonstrate its utility, we employ our extended grammar to recreate interactive examples, showcasing its ability to represent complex visual designs. Comprehensive examples and an interactive editor are available at 3d.gosling-lang.org.
- Open Access
- 6 November 2025
- Published by Public Library of Science (PLoS) in PLOS Computational Biology
- Vol. 21 (11) , e1013657
Ten simple rules for making biomedical data resources accessible
- 6 November 2025
- Published by Public Library of Science (PLoS) in PLOS Computational Biology
- Vol. 21 (11) , e1013657
Open AccessTen simple rules for making biomedical data resources accessible
- Open Access
- 25 July 2025
- Published by Center for Open Science in OSF Preprints
Ten Simple Rules for Making Biomedical Data Resources Accessible
As computational biologists, we share a responsibility to make our data and tools usable and accessible to everyone in our community. However, without digital accessibility, biomedical data resources that are essential for data-driven research (e.g., data portals, visualization tools) will exclude people with disabilities from entering the scientific workforce. This is evidenced by the significant gap between the number of people with disabilities in the scientific workforce (9%) and the general population (26%). Our recently published large-scale evaluation of essential biomedical data resources found that over 99% of them are largely inaccessible to people with disabilities.We present ten simple rules for making web-based biomedical data resources accessible for people with disabilities. Over the last few years, we have conducted a comprehensive analysis of the accessibility landscape of biomedical data resources and developed multiple tools to enhance their accessibility. Our ten simple rules are a combination of lessons we have learned along the way, as well as practical advice for addressing the most common issues that we have found. These rules are actionable recommendations that help computational biologists, bioinformaticians, and others in the field to quickly and directly improve the accessibility of their biomedical data resources.
- 25 July 2025
- Published by Center for Open Science in OSF Preprints
Open AccessTen Simple Rules for Making Biomedical Data Resources Accessible
As computational biologists, we share a responsibility to make our data and tools usable and accessible to everyone in our community. However, without digital accessibility, biomedical data resources that are essential for data-driven research (e.g., data portals, visualization tools) will exclude people with disabilities from entering the scientific workforce. This is evidenced by the significant gap between the number of people with disabilities in the scientific workforce (9%) and the general population (26%). Our recently published large-scale evaluation of essential biomedical data resources found that over 99% of them are largely inaccessible to people with disabilities.We present ten simple rules for making web-based biomedical data resources accessible for people with disabilities. Over the last few years, we have conducted a comprehensive analysis of the accessibility landscape of biomedical data resources and developed multiple tools to enhance their accessibility. Our ten simple rules are a combination of lessons we have learned along the way, as well as practical advice for addressing the most common issues that we have found. These rules are actionable recommendations that help computational biologists, bioinformaticians, and others in the field to quickly and directly improve the accessibility of their biomedical data resources.
- Open Access
- 8 May 2025
- Published by Center for Open Science in OSF Preprints
Design Space and Declarative Grammar for 3D Genomic Data Visualization
Various computational approaches predict chromatin structure, yielding concrete models that position genomic loci in physical space and help reveal genome organization and function. While prior visualization research has explored data and task abstractions for genomics, the design space for depicting these three-dimensional (3D) genome models---and associated genome-mapped data---remains unclear. In this paper, we investigate the visualization of genomic data with a spatial component. First, we systematically survey how 3D genome models are used and depicted in computational biology. We analyze over 300 papers with figures that visualize 3D genomic data and categorize the methods for visual representation. From this survey, we derive a design space for visualizing 3D genome data, identifying common patterns and key properties such as representation, visual channels, and composition. We position these findings within an existing genomics visualization taxonomy, refining and extending existing classifications. Second, we augment Gosling, a declarative visualization grammar for genomics, to support 3D genomic data. Our integration enables expressive authoring of visualizations that connect traditional genome-mapped information with 3D genome models, emphasizing their spatial characteristics. To demonstrate its utility, we employ our extended grammar to recreate interactive examples, showcasing its ability to represent complex visual designs. Comprehensive examples and an interactive editor are available at 3d.gosling-lang.org.
- 8 May 2025
- Published by Center for Open Science in OSF Preprints
Open AccessDesign Space and Declarative Grammar for 3D Genomic Data Visualization
Various computational approaches predict chromatin structure, yielding concrete models that position genomic loci in physical space and help reveal genome organization and function. While prior visualization research has explored data and task abstractions for genomics, the design space for depicting these three-dimensional (3D) genome models---and associated genome-mapped data---remains unclear. In this paper, we investigate the visualization of genomic data with a spatial component. First, we systematically survey how 3D genome models are used and depicted in computational biology. We analyze over 300 papers with figures that visualize 3D genomic data and categorize the methods for visual representation. From this survey, we derive a design space for visualizing 3D genome data, identifying common patterns and key properties such as representation, visual channels, and composition. We position these findings within an existing genomics visualization taxonomy, refining and extending existing classifications. Second, we augment Gosling, a declarative visualization grammar for genomics, to support 3D genomic data. Our integration enables expressive authoring of visualizations that connect traditional genome-mapped information with 3D genome models, emphasizing their spatial characteristics. To demonstrate its utility, we employ our extended grammar to recreate interactive examples, showcasing its ability to represent complex visual designs. Comprehensive examples and an interactive editor are available at 3d.gosling-lang.org.
- Open Access
The mitochondrial NAD transporter SLC25A51 is a modulator of beta cell senescence and type 2 diabetes
: Nicotinamide adenine dinucleotide (NAD + ) is an essential redox cofactor and signaling molecule linked to age-dependent metabolic decline, with its compartmentalization regulated by the mitochondrial carrier SLC25A51. The mechanisms contributing to declining NAD + levels during aging and the consequences of altered NAD + homeostasis across tissues are poorly understood. Here, we show that SLC25A51 is upregulated in aging and aging-associated conditions, particularly in senescent cells. In a mouse model of beta-cell senescence, upregulated SLC25A51 was associated with beta-cell identity loss, senescence progression, and a reduced NAD + /NADH ratio. SLC25A51 was elevated following p16 INK4a -, replicative-, irradiation-, and H 2 O 2 -induced senescence, with NRF2 implicated as a potential transcriptional regulator. Overexpression of SLC25A51, but not a transport-dead mutant, induced senescence factors, while its deletion prevented this effect. Beta-cell-specific deletion of SLC25A51 lowered p16 INK4a levels in pancreatic islets, circulating insulin, and glucose levels, improving insulin sensitivity and indicating its role in cellular senescence and the metabolic control of beta-cell function.
Open AccessThe mitochondrial NAD transporter SLC25A51 is a modulator of beta cell senescence and type 2 diabetes
: Nicotinamide adenine dinucleotide (NAD + ) is an essential redox cofactor and signaling molecule linked to age-dependent metabolic decline, with its compartmentalization regulated by the mitochondrial carrier SLC25A51. The mechanisms contributing to declining NAD + levels during aging and the consequences of altered NAD + homeostasis across tissues are poorly understood. Here, we show that SLC25A51 is upregulated in aging and aging-associated conditions, particularly in senescent cells. In a mouse model of beta-cell senescence, upregulated SLC25A51 was associated with beta-cell identity loss, senescence progression, and a reduced NAD + /NADH ratio. SLC25A51 was elevated following p16 INK4a -, replicative-, irradiation-, and H 2 O 2 -induced senescence, with NRF2 implicated as a potential transcriptional regulator. Overexpression of SLC25A51, but not a transport-dead mutant, induced senescence factors, while its deletion prevented this effect. Beta-cell-specific deletion of SLC25A51 lowered p16 INK4a levels in pancreatic islets, circulating insulin, and glucose levels, improving insulin sensitivity and indicating its role in cellular senescence and the metabolic control of beta-cell function.
- Open Access
- 17 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Visualization and Computer Graphics
- Vol. 31 (1) , 1180-1190
Understanding Visualization Authoring Techniques for Genomics Data in the Context of Personas and Tasks
Genomics experts rely on visualization to extract and share insights from complex and large-scale datasets. Beyond off-the-shelf tools for data exploration, there is an increasing need for platforms that aid experts in authoring customized visualizations for both exploration and communication of insights. A variety of interactive techniques have been proposed for authoring data visualizations, such as template editing, shelf configuration, natural language input, and code editors. However, it remains unclear how genomics experts create visualizations and which techniques best support their visualization tasks and needs. To address this gap, we conducted two user studies with genomics researchers: (1) semi-structured interviews (n=20) to identify the tasks, user contexts, and current visualization authoring techniques and (2) an exploratory study (n=13) using visual probes to elicit users' intents and desired techniques when creating visualizations. Our contributions include (1) a characterization of how visualization authoring is currently utilized in genomics visualization, identifying limitations and benefits in light of common criteria for authoring tools, and (2) generalizable design implications for genomics visualization authoring tools based on our findings on task- and user-specific usefulness of authoring techniques. All supplemental materials are available at https://osf.io/bdj4v/
- 17 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Visualization and Computer Graphics
- Vol. 31 (1) , 1180-1190
Open AccessUnderstanding Visualization Authoring Techniques for Genomics Data in the Context of Personas and Tasks
Genomics experts rely on visualization to extract and share insights from complex and large-scale datasets. Beyond off-the-shelf tools for data exploration, there is an increasing need for platforms that aid experts in authoring customized visualizations for both exploration and communication of insights. A variety of interactive techniques have been proposed for authoring data visualizations, such as template editing, shelf configuration, natural language input, and code editors. However, it remains unclear how genomics experts create visualizations and which techniques best support their visualization tasks and needs. To address this gap, we conducted two user studies with genomics researchers: (1) semi-structured interviews (n=20) to identify the tasks, user contexts, and current visualization authoring techniques and (2) an exploratory study (n=13) using visual probes to elicit users' intents and desired techniques when creating visualizations. Our contributions include (1) a characterization of how visualization authoring is currently utilized in genomics visualization, identifying limitations and benefits in light of common criteria for authoring tools, and (2) generalizable design implications for genomics visualization authoring tools based on our findings on task- and user-specific usefulness of authoring techniques. All supplemental materials are available at https://osf.io/bdj4v/
- Open Access
- 10 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Visualization and Computer Graphics
- Vol. 31 (1) , 459-469
Learnable and Expressive Visualization Authoring Through Blended Interfaces
A wide range of visualization authoring interfaces enable the creation of highly customized visualizations. However, prioritizing expressiveness often impedes the learnability of the authoring interface. The diversity of users, such as varying computational skills and prior experiences in user interfaces, makes it even more challenging for a single authoring interface to satisfy the needs of a broad audience. In this paper, we introduce a framework to balance learnability and expressivity in a visualization authoring system. Adopting insights from learnability studies, such as multimodal interaction and visualization literacy, we explore the design space of blending multiple visualization authoring interfaces for supporting authoring tasks in a complementary and flexible manner. To evaluate the effectiveness of blending interfaces, we implemented a proof-of-concept system, Blace, that combines four common visualization authoring interfaces–template-based, shelf configuration, natural language, and code editor–that are tightly linked to one another to help users easily relate unfamiliar interfaces to more familiar ones. Using the system, we conducted a user study with 12 domain experts who regularly visualize genomics data as part of their analysis workflow. Participants with varied visualization and programming backgrounds were able to successfully reproduce unfamiliar visualization examples without a guided tutorial in the study. Feedback from a post-study qualitative questionnaire further suggests that blending interfaces enabled participants to learn the system easily and assisted them in confidently editing unfamiliar visualization grammar in the code editor, enabling expressive customization. Reflecting on our study results and the design of our system, we discuss the different interaction patterns that we identified and design implications for blending visualization authoring interfaces.
- 10 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Visualization and Computer Graphics
- Vol. 31 (1) , 459-469
Open AccessLearnable and Expressive Visualization Authoring Through Blended Interfaces
A wide range of visualization authoring interfaces enable the creation of highly customized visualizations. However, prioritizing expressiveness often impedes the learnability of the authoring interface. The diversity of users, such as varying computational skills and prior experiences in user interfaces, makes it even more challenging for a single authoring interface to satisfy the needs of a broad audience. In this paper, we introduce a framework to balance learnability and expressivity in a visualization authoring system. Adopting insights from learnability studies, such as multimodal interaction and visualization literacy, we explore the design space of blending multiple visualization authoring interfaces for supporting authoring tasks in a complementary and flexible manner. To evaluate the effectiveness of blending interfaces, we implemented a proof-of-concept system, Blace, that combines four common visualization authoring interfaces–template-based, shelf configuration, natural language, and code editor–that are tightly linked to one another to help users easily relate unfamiliar interfaces to more familiar ones. Using the system, we conducted a user study with 12 domain experts who regularly visualize genomics data as part of their analysis workflow. Participants with varied visualization and programming backgrounds were able to successfully reproduce unfamiliar visualization examples without a guided tutorial in the study. Feedback from a post-study qualitative questionnaire further suggests that blending interfaces enabled participants to learn the system easily and assisted them in confidently editing unfamiliar visualization grammar in the code editor, enabling expressive customization. Reflecting on our study results and the design of our system, we discuss the different interaction patterns that we identified and design implications for blending visualization authoring interfaces.