Single-cell data is high-dimensional data that can address unanswered questions in different biological areas. Visualization tools play a pivotal role in navigating single-cell datasets, offering structured user experiences and pathways through their layouts, features, and functionalities. It’s roughly been a year since we released our blog on- Comparison of Single Cell Data Visualization Tools. Since the first blog was well-received, we decided to follow suit and explore more single-cell data visualization tools.

We explored over 30 different applications,

‍and evaluated them based on features, popularity (as a proxy to a number of users), development status, and issue resolution. We are also not including portals or stand-alone applications where a preselected limited set of datasets by the developer/host data are available and importing or hosting by the user is not supported. Within these criteria, we present 5 applications in this blog, worthy of a single-cell data explorer’s attention. Dig in!

Azimuth

Azimuth, developed by Satija Lab (Pioneer behind many single-cell tools, the most famous of them being Seurat) as a part of the NIH Human Biomolecular Atlas Project (HuBMAP), leverages existing single-cell reference datasets to automate the processing, annotation, analysis, and automation of the scRNASeq and scATACSeq data. It is one of the most popular applications as it supports Seurat-based analysis in combination with a visual interface. While 13 references are released by HuBMAP to support Azimuth and a few other projects, more references are in development. It also supports using a custom reference created separately.

UCSC Cell Browser

UCSC Cell Browser is a lightweight exploration tool for single-cell data. It provides a static, precomputed, fast-responding layout for easy exploration of single-cell data. There are exporters built within Seurat, Scanpy, and other tools for exporting a dataset to create a UCSC Cell Browser that can be hosted on a server privately or on the web. It relies on pre-computed values from the single-cell datasets for providing fast responsive datasets. A tsne/umap-centered layout provides a great medium to overlay gene expression and metadata over the cells.

Cellenium

Cellenium developed by Bayer is an application designed for scaleability. The application is powered by a PostgreSQL database backend queriable using the GraphQL APIs which makes the application very scalable. Cellenium provides an interface for single dataset exploration, multi-dataset exploration, and marker gene exploration using a searchable layout. The search is further backed by ontologies for cell type, disease, tissue, and organism from MeSH, Cell Ontology, and NCIT which makes the exploration more FAIR.  

Asc-Seurat

Asc-Seurat, pronounced as ‘Ask-Seurat' is a Seurat, Dynverse and Biomart-based application hosted on a shiny framework. Asc-Seurat provides end-to-end processing for single-cell data imported from 10x format coupled with trajectory analysis using dynverse. We found the customizability of plots as a great feature of the application. It’s a simple application for the exploration of single-cell data starting from raw files. A minimal downside is no active developments on Github in over a year and no active issue resolution in over 2 years.

Scala

Scala is a recently released (year 2023) application that can be a great alternative to Asc-Seurat. It is a multi-modal application that offers end-to-end processing and analysis of both scRNASeq and scATACSeq data. It also supports great features like trajectory analysis, ligand-receptor, pair analysis, and gene regulatory network analysis. While Asc-Seurat has not been actively developed for a while, Scala is very new and has no active user testing data or issues present on its GitHub repository. We are hopeful that the application which boasts great features will gain some traction and user testing over time.

‍

‍

Visualization applications for single-cell data is an interesting problem that requires solving technical challenges while maintaining scientific reliability to provide a user flow for efficient exploration. Since the data itself is very high dimensional, there is no one-size-fits-all application for everyone. In our blog, we covered different applications, some designed for speed and response, some designed for multiple downstream analysis and some supporting multiple data types.

Polly by Elucidata supports the exploration of single-cell data using CellxGene by default and also provides solutions to host public applications and build custom applications. The applications are a great powerhouse on top of the curated and harmonized single-cell data hosted on Polly.

Connect with us or reach out to us at info@elucidata.io to learn more.