🧬 LIVE-Bio

Background

There are many interactive visualization tools in Biology, including a growing number of beautiful web-based tools. Most of the tools focus on visualizing a single type of data, requiring researchers who want to interrogate and learn from multiple data types in their work to switch between multiple tools to do their science. This workflow, which may involve manually uploading data to web services and then downloading results/subsets for use in other tools, reduces reproducibility, and makes exploring multiple hypotheses cumbersome.

Special expertise is often required to access, analyze, and visualize especially large datasets. The resulting asynchronous interactions between researchers and bioinformatics specialists reduces the opportunity for data exploration, impeding insight and discovery.

Current Capabilities

The glue genes project, developed as a partnership between glue solutions, inc. and The Jackson Laboratory, is a desktop-based prototype for LIVE-Bio, demonstrating how multiple data modalities and visualization types can be combined to provide a more seamless environment for exploratory data analysis.

The video from which the screen shot here is drawn gives an example exploration of spatial transcriptomics data from the mouse brain using glue genes.

Vitessce integrates multiple interactive viewers for the exploration of spatial single cell experiments. Vitessce also showcases the flexibility that modern web-development provides, as it can be used as a standalone web application, embedded within another web application, or used as a Python jupyter widget.

The Future

The LIVE-Bio environment will allow users to explore and visualize a wide range of biological datasets in an interactive environment. Here is a list of core requirements for LIVE-Bio functions, along with some existing example viewers. Given the inherently modular nature of the LIVE approach, users will be able to choose pre-configured sets of interlinked tools, or create their own environments.

Core functionality requirements for LIVE-Bio

display very large multi-resolution images, including images hosted remotely (e.g. image viewer in glue genes or Viv)
viewer for genome browser tracks (such as Gosling or igv.js)
3D viewers for both point and volumetric data (such as ITK/VTK, Ipyvolume, Vispy)
network viewer for displaying complex relationships between biological components (such as Cytoscape.js)
interactions with online databases (e.g. HumanBase, GeneWeaver, Enrichr) to enrich local datasets and (where appropriate) contribute information back to collaborative databases
general-purpose visualization tools such as scatterplots, histograms, heatmaps, tables, and other statistical plots (such as those included in glue genes).