No.086 Web-based Molecular Graphics

Icon

NII Shonan Meeting Seminar 086

Description of the meeting

Summary

Molecular graphics is a well established discipline used in many scientific fields (e.g., life sciences, biomedical research, chemistry, material sciences), including commercial research (e.g., pharmaceutical industry). A wealth of methods and tools have been developed, and are still actively developed – however until recently activity has focused on stand-alone software. This meeting will explore an emerging new frontier for molecular graphics, namely deployment in a server-client environment, which promises to make molecular structural information much more easily accessible to scientists. In addition, the recent hybrid approaches have revealed structures, functions and dynamics of very large molecules and cellular machineries, which are displayed as both atomic structures and molecular/cellular images, greatly extending the roles of molecular graphics. This meeting will bring together key players in this emerging field to explore the challenges and opportunities raised by web-based molecular graphics and new applications.

Background

A rapidly growing wealth of molecular structure data have been derived from experimental methods (e.g., X-ray crystallography, NMR spectroscopy, electron microscopy) and are collected in the PDB (Protein Data Bank). In addition computational methods, such as modeling and molecular dynamics, continue to expand upon the PDB data. In combination, all these types of structural data provide scientists today with unprecedented insight into a wide range of molecular phenomena, for example protein function, biological membrane transport, or material properties.

The amount of data is steadily growing in size as well as in complexity: this includes more PDB entries, larger structures (big complexes) and time-resolved data, such as molecular dynamics trajectories. In the near future, these data may increase in scale to include, for example, complex cellular machines and whole organelles or cells. A key challenge is to make all of these data accessible to scientists, in ways that enable interpretation and the generation of new research insights.

Fortunately, nowadays almost all scientists – even students – have easy access to computer graphics capabilities powerful enough to show a wide variety of molecular structures. With modern bandwidth speeds, it is even feasible to stream complex, interactive molecular graphics into simple devices such as an iPad or iPhone, e.g., by server-side GPU-rendering. Thus, it is now possible to depict and explore 3D models in regular browsers, such as Mozilla Firefox, Internet Explorer or Google Chrome. This allows to bring 3D molecular visualisation tools to non-expert end users without the need to install specialised software. However, in times of a steadily growing amount of data, improved visualization methods and tools are urgently needed, which can make the more complex data accessible and interpretable.

 

Aims of the Seminar

We believe that web molecular graphics is likely to soon become very widely adopted; therefore, we feel it is very timely to host a Shonan meeting on this topic. Overall, the seminar aims to help give this emerging field direction and clarity by streamlining global efforts.

We aim to discuss strategies for effectively managing very large molecular structures (e.g., millions of atoms or more). In some cases (e.g., whole organelles, entire genomes, or materials, such as metal-organic frameworks), the molecular assembly becomes so complex that multi-scale, hierarchical representations are required, e.g., changing representation automatically at different scales, somewhat analogously to Google Maps. Similar strategies can also be needed for large molecular dynamic simulations. These cases can be challenging for web molecular graphics, which is limited by network bandwidth.

In addition, many of the scientific problems being investigated cannot be solved with molecular graphics alone; instead, data on molecular structures needs to be integrated with other kinds of data that uses other visualization paradigms, e.g., 1D sequence-based features, 2D interrelationships shown on network graphs, etc. Therefore, approaches are required that are compatible to different data resources. JavaScript, JSON, and REST has emerged as defacto standards as has WebGL for 3D visualization. The seminar will discuss pros and cons of these standards and aims to define desirable interaction paradigms and communication methods.