Polyphony: an Interactive Transfer Learning Framework for Single-Cell Data Analysis

Cheng, Furui; Keller, Mark S; Qu, Huamin; Gehlenborg, Nils; Wang, Qianwen

doi:10.1109/tvcg.2022.3209408

Cited by 6 publications

(5 citation statements)

References 66 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We talked about the Transmorph framework that articulates computational blocks to conceive HI pipelines, but this is not the only framework that exists which is related to data integration. We can cite MUON (Bredikhin et al, 2022), which facilitates the handling of data consisting of different modalities, Polyphony (Cheng et al, 2022), which carries out transfer learning across datasets by leveraging data integration algorithms, or SinCast (Deng et al, 2022) which is specialized in cell type inference by mapping a query onto an atlas.…”

Section: Discussionmentioning

confidence: 99%

Omics data integration in computational biology viewed through the prism of machine learning paradigms

Fouché

Zinovyev²

2023

Front. Bioinform.

View full text Add to dashboard Cite

Important quantities of biological data can today be acquired to characterize cell types and states, from various sources and using a wide diversity of methods, providing scientists with more and more information to answer challenging biological questions. Unfortunately, working with this amount of data comes at the price of ever-increasing data complexity. This is caused by the multiplication of data types and batch effects, which hinders the joint usage of all available data within common analyses. Data integration describes a set of tasks geared towards embedding several datasets of different origins or modalities into a joint representation that can then be used to carry out downstream analyses. In the last decade, dozens of methods have been proposed to tackle the different facets of the data integration problem, relying on various paradigms. This review introduces the most common data types encountered in computational biology and provides systematic definitions of the data integration problems. We then present how machine learning innovations were leveraged to build effective data integration algorithms, that are widely used today by computational biologists. We discuss the current state of data integration and important pitfalls to consider when working with data integration tools. We eventually detail a set of challenges the field will have to overcome in the coming years.

show abstract

Section: Discussionmentioning

confidence: 99%

Omics data integration in computational biology viewed through the prism of machine learning paradigms

Fouché

Zinovyev²

2023

Front. Bioinform.

View full text Add to dashboard Cite

show abstract

“…Although insightful, previous studies summarize tasks merely based on interviewing a limited number of domain experts [4,8,11] or reviewing literature from the visualization community [21,27,35]. Building on previous studies, this study presents a more comprehensive characterization of the tasks involved in HD data visualization in the wild beyond the visualization community.…”

Section: Understanding the Usage Of Hd Data Visualizationmentioning

confidence: 99%

“…Visualization has been widely employed for HD data analysis [21,27,40], including visualizing multidimensional data (e.g., parallel coordinates, scatterplot matrices) and visualizing more interpretable lowdimensional data derived from the original HD data (e.g., embedding visualization [11,50]).…”

Section: Understanding the Usage Of Hd Data Visualizationmentioning

confidence: 99%

What are People Talking about in High-Dimension Data Visualization? LLM-supported Analysis of Domain Literature

Wang,

liu,

Gehlenborg

2024

Preprint

View full text Add to dashboard Cite

Visualizing high-dimensional (HD) data is a common yet challenging task in various application domains. Previous surveys on HD visualization are conducted either in the visualization community or through the interview of a limited number of domain users. A comprehensive understanding of the usage of HD visualizations in the wild is missing. To fill this gap, we analyzed more than 1,000 papers from one representative domain (single-cell transcriptomics) that extensively employed HD data visualizations. To effectively analyze this extensive corpus filled with highly domain-specific terminologies, we propose a pipeline to effectively collaborate with an LLM annotator on interpreting and summarizing the usage of HD visualizations in the collected papers.This pipeline includes machine learning techniques for figure detection, traditional NLP methods for text cleaning, and LLM prompt engineering for nuanced interpretation.With this pipeline, we categorized HD visualization based on how users referred to and mentioned these visualizations in their papers.We then discussed representative visualizations for each category, as well as current practices and potential wrong uses.These analyses can assist the visualization community in designing and evaluating future HD visualizations.

show abstract

“…Emblaze [57] offers animated scatter plots for interactive comparison within Jupyter notebooks but lacks support for distinct datasets with corresponding classes. Polyphony [9] merges interactive visualization of single-cell latent representations to aid human-driven cell annotation about an "anchor" embedding. However, its primary design aids cell type annotation, leaving a gap in tools addressing comparisons of annotated cell types between groups.…”

Section: Techniques For Comparing Embedding Visualizationsmentioning

confidence: 99%

A General Framework for Comparing Embedding Visualizations Across Class-Label Hierarchies

Manz,

Lekschas,

Greene

et al. 2024

Preprint

View full text Add to dashboard Cite

Projecting high-dimensional vectors into two dimensions for visualization, known as embedding visualization, facilitates perceptual reasoning and interpretation. Comparison of multiple embedding visualizations drives decision-making in many domains, but conventional comparison methods are limited by a reliance on direct point correspondences. This requirement precludes embedding comparisons without point correspondences, such as two different datasets of annotated images, and fails to capture meaningful higher-level relationships among point groups. To address these shortcomings, we propose a general framework to compare embedding visualizations based on shared class labels rather than individual points. Our approach partitions points into regions corresponding to three key class concepts---confusion, neighborhood, and relative size---to characterize intra- and inter-class relationships. Informed by a preliminary user study, we realize an implementation of our framework using perceptual neighborhood graphs to define these regions and introduce metrics to quantify each concept. We demonstrate the generality of our framework with use cases from machine learning and single-cell biology, highlighting our metrics' ability to draw insightful comparisons across label hierarchies. To assess the effectiveness of our approach, we conducted a user study with five machine learning researchers and six single-cell biologists using an interactive and scalable prototype developed in Python and Rust. Our metrics enable more structured comparison through visual guidance and increased participants’ confidence in their findings.

show abstract

Polyphony: an Interactive Transfer Learning Framework for Single-Cell Data Analysis

Cited by 6 publications

References 66 publications

Omics data integration in computational biology viewed through the prism of machine learning paradigms

Omics data integration in computational biology viewed through the prism of machine learning paradigms

What are People Talking about in High-Dimension Data Visualization? LLM-supported Analysis of Domain Literature

A General Framework for Comparing Embedding Visualizations Across Class-Label Hierarchies

Contact Info

Product

Resources

About