-
Notifications
You must be signed in to change notification settings - Fork 7
HiView User Guide
This is the documentation for HiView classic. Click here for DDRAM documenation
Introduction
Why using HiView
How to use HiView
How to prepare your data for HiView
References
The HiView is an interactive web-based visualization tool of hierarchical models.
Hierarchical models provide an effective abstraction of data in many scientific fields. In systems biology, such models are based on grouping biological entities (e.g. genes, proteins, cells, species) into systems at multiple scales. The main feature of hierarchical models is the nesting of systems, i.e. a small system can be seen as a subcomponent of one or more large systems.
Hierarchical models often contain pleiotropic systems, i.e. a system can be a child of multiple parent systems. As a result, models take the form of directed acyclic graphs (DAGs), making it challenging to clearly visualize models with traditional formats like dendrograms. HiView allows the user to interactively zoom across many scales in the model, from an expansive view of the entire hierarchy to more focused views of particular systems; HiView also includes convenient searching functions, which makes it easier to locate any entity of interest in the model.
HiView is tightly integrated with the research and tools in network biology. In HiView, each system can be coupled with network data to show the interactions between members of the system. HiView also provides APIs to Cytoscape, NDEx, GeneCards, and Enrichr, making it easy to query and expand the information of genes and systems of interest.
The HiView interface includes the following components:
In the model view, HiView uses a circle packing view from D3.js to represent the nested systems. In the viewport upon instantiating a session, the outer circle is the root of the model (a system containing all subsystems and genes), and the inner circles represent all the 1st level subsystems. The name of each system is shown in each circle.
Note: here we use "genes" as a general term referring to all terminal entities (i.e. those do not have further substructures) in a model; since the design of HiView was mainly motivated by visualizing hierarchical models of genes, and there are functionalities developed specifically for genes (such as APIs for NDEx-iQuery, GeneCards, and Enrichr). Nonetheless, HiView does not require the terminal entities to be genes.
The model can be browsed with intuitive mouse operations:
- Double-click on a circle: expand a system and display all the subsystems and genes in the next level. Also, it will load a network (when available) in the Data View supporting the interactions within the system.
- Double-click on a gene will display the information of the gene, such as aliases, gene IDs, and associated GO terms.
- Hold and drag: move the viewport
- Scroll the mouse: zoom in and out for the viewport
- Hovering on a circle: when the viewport focuses on a system and if there is a network in the data view, hovering mouse on the circle of a subsystem will highlight the genes belonging to this subsystem in the network.
Note: if a system/gene is a child of multiple parent systems, it will appear to be nested inside the circles of all its parent systems.
On the top, the user can browse the interaction network supporting a system selected in the Model View. The network can contain multiple types of interactions, and the user can turn particular types on and off with the toggles below the network window. Different types of interactions are presented in different colors. The network can be downloaded via the "Open in Cytoscape" button.
At the bottom part of the data view (the "Subsystem Detail" tab), the user can see the metadata of a system (number of genes, associated diseases, etc.). The available information is defined by the model creator. The user can also see a full list of genes (the "Assigned Genes" tab).
The search panel can be used to search and locate genes or systems in the model. There are three options of search:
- exact match (for genes): searching "MYH1" will only return results of "MYH1"
- prefix match (for genes): searching "MYH1" may also return results of "MYH10"
- fuzzy match (for systems): results will contain all systems for which the names contain the queried string.
The control panel can be opened via the hamburger button on the top left. It allows the user to change colors, change depths of circle packing rendering, and turn on/off the enrichment functionality.
When the enrichment function is turned on (in the control panel), double-clicking a system will run gene set enrichment against a few major databases (GO, KEGG, etc.), and return a list of gene sets significantly enriching the genes in the system.
This function is powered by the Enrichr API.
The user can create a customized hierarchical model and upload it to HiView via the DDOT package. Please see the tutorial we provide (which requires DDOT v1.0.1+ and HiView 2.6+)
If you use HiView in research we wish you cite either of the following papers:
Zheng, F. et al. Convergence of cancer mutations on a hierarchy of protein systems. Under review.
Yu, M. K. et al. DDOT: A Swiss Army Knife for Investigating Data-Driven Biological Ontologies. Cell Syst 8, 267–273.e3 (2019)