Date: 28th April 2021
Complex networks are studied across many fields of science, and are powerful way to analyse complex and diverse data and systems. However, with the ever increasing size of datasets made possible with the next generation of advanced technologies, the ability to visualise these large networks is limiting. Whilst, conventional computer programs use to offer us an easy way to see and interact with simple networks, they quickly reach their limit and interacting with the network via a screen becomes difficult. Now, scientists have developed a Virtual Reality (VR) platform for exploring huge amounts of data and their complex interplay in a uniquely intuitive fashion. They use the platform as a proof of concept to explore genome-scale molecular networks to identify genes associated with rare diseases and to link them to how they contribute to disease development.
In biology and medicine, networks are being used for numerous applications for disentangling the enormous complexity within and across different levels of biological organisation. For example, in molecular interaction networks we can exploit the data to extract which genes are involved in healthy and disease states, can assign genes within a dense node to specific biological functions, or identify disease-associated processes via distinct connection patterns.
However, currently the full potential of these networks is constrained by the way we visualise and interact with them. The true multi-scale nature of the biological processes being studied can only be fully appreciated when the entire range of the global network structure is available – from zooming in to a local view, then expanding out to see the global picture, the ability to scrutinise and analyse the entire landscape would be hugely beneficial but requires technology beyond a computer screen.
Now, scientists led by Jörg Menche at the CeMM Research Center for Molecular Medicine of the Austrian Academy of Sciences, and the University of Vienna, Austria, have used immersive Virtual Reality technology to create an exploration platform, called VRNetzer, to visualise and interact with large and complex data. They show how the platform can be used to investigate gene variants in the context of a molecular interaction network to identify variant responsible for severe genetic diseases.
The VRNetzer platform was designed in a modular fashion, allowing the user to customise and extend the components for data visualisation, data analysis, and data input separately. It consisted of 5 key modules ; a data module, analytics module, the user interface (UI) module which served as a communication layer, and the frontend of the system which consisted of the VR interface module with an additional web browser-based interface web module which then completed the platform.
As an input, the team used a molecular network curated from the literature, as well as widely used gene annotation data - ranging from molecular functions to disease associations. The resulting genome-scale molecular network represented the human interactome (the set of protein–protein interactions (the interactome) that occur in human cells) and consisted of around 16,000 proteins as nodes, and around 300,000 physical interactions as links. This was the first time the entirety of protein interactions had been made visible.
The team had drawn on technology normally used in the development of 3D computer games for their VR interface. This allowed basic navigation of the network, free movement within the surrounding network, rotation, and translation of the network with full six degree of freedom (forward/backwards/up/down/left/right), as well as arbitrary scale of the network size. This led to a seamless transition between global network views and zooming into to close-up inspection of local node clusters or even individual nodes. The user could explore the data in a uniquely intuitive fashion, without the need of specialised knowledge in VR, allowing them to quickly locate nodes and node clusters, and identify their immediate and broader landscape in different functional contexts.
They used the network to identified connection patterns between different protein complexes in the human body and linked them to their biological functions. In addition, the scientists used the global databases to identify specific protein complexes associated with a particular disease. For example they considered a patient suffering from severe combined immune deficiency with an unknown genetic cause, and using the different features of the platform generated and identified potential candidates.
The team here have demonstrated the power of VR as the basis of a new approach to visualise the ever increasing size and complexity of data-driven methodologies such as networks. By exploiting human capabilities such as intuition and creativity, VR offers an exciting platform to interact with data on a scale previously not possible. They present identification and functional interpretation of a genomic variant responsible for a rare disease. Highlighting how the platform can smoothly navigate computational tools, diverse data resources and expert interpretation of complex biological relationships.
The team hope that their open-source code will be used within a broad community, allowing other developers to use VR for analysing a range of scientific data, and will extend outside the field of biology.
VR offers a truly unique look into our own bodies and biology. Here, this work offers the first steps into understanding large and complex data and building the next generation of data exploration platforms. We recently reported vLUME (visualization of the local universe in a micro environment), an immersive virtual reality (VR)-based visualisation software package purposefully designed to render large 3D- super-resolution microscopy datasets. In essence, allowing us to walk inside our own cells. Together, these types of platforms offer users an entirely different perspective of their work, and will accelerate new discoveries both in healthy and disease states. With VR technologies also evolving such that a sense of touch can now be added to the VR experience, it will be fascinating to see the applications that will be enhanced in the near future by early adoption of these technologies.
For more information please see the press release from CeMM
Pirch, S., Müller, F., Iofinova, E., Pazmandi, J., Hütter, C.V.R., Chiettini, M., Sin, C., Boztug, K., Podkosova, I., Kaufmann, H., et al. (2021). The VRNetzer platform enables interactive network analysis in Virtual Reality. Nature Communications 12, 2432.
https://doi.org/10.1038/s41467-021-22570-w
