On the weekend of July 28th-30th, about 150 NMR (nuclear magnetic resonance) scientists from the UK and Europe (and a few from the US) gathered in Ambleside for the Sixth Annual Collaborative Computing Project for NMR (CCPN) Meeting. The topic of the meeting was "Efficient and Rapid Structure Determination by NMR", and it included presentations by twelve NMR scientists, a demonstration of the CCPN software, and about thirty research posters on various NMR talks.
Note: for more background on NMR, check out my previous post on the topic.
Biomolecular NMR is a rapidly growing field, one that has attracted an increasing number of researchers as its tools and techniques become more advanced and sophisticated. Due to this growth and the relative youth of the field, there are very few, if any, standardized methods in NMR, with researchers using a host of different means of acquiring, processing, and analyzing data, means that are only semi-compatible with each other.
To address the need for more continuity and connectivity within the biomolecular community in the UK, CCPN was set up in 1999, modeled after the longstanding CCP4 project for x-ray crystallography. Although scientists in the field of protein x-ray crystallography still use a variety of programs, the methods have at least become more standardized and the data more transferable between platforms.
The goals of CCPN are:
- to promote the exchange of ideas through the arrangement of meetings and workshops;
- to establish a universal data model for harvesting and exchange of data between different NMR related software packages, and develop associated subroutine libraries;
- to develop software packages using the common data model to perform standard functions required during the determination of macromolecular structures by NMR; and,
- to promote the use of the common data model and to diffuse NMR related software made by third parties, especially those that make use of the data model, within the NMR community
In short, while CCPN develops and pushes its own software, it is mainly geared toward allowing protein NMR researchers to continue to use their preferred programs by improving the transferability of data between different programs. In the age of mass customization, this philosophy certainly appeals to the prevailing sentiment of the day, allowing members of the NMR community to continue to use their preferred methods, hopefully more efficiently, without restricting the choices available
Central to the work of CCPN is the organizing of annual conferences on various topics in NMR. Although the theme of this year's conference was "Efficient and Rapid Structure Determination by NMR", the presentations covered a variety of NMR topics, from general textbook topics to the latest developments in the field.
I'm still fairly new to the field of protein NMR, but high-speed structure determination is an area that I'm particularly unfamiliar with. Therefore, the conference was particularly valuable to me as a quick introduction to the possibilities available as I look to begin some structural projects in the near future.
Solving a protein structure by NMR is an involved and time-consuming process. Although producing the sample needed for the experiments can be one of the most difficult aspects of determining a structure, there are several roadblocks within the actual structure determination process, including acquiring the necessary spectra, assigning the peaks in the spectra to specific atoms in the protein, and then using the acquired data to solve the structure of the protein.
One presentation by Bernhard Brutscher of the Institut de Biologie Structurale in Grenoble, France, discussed a variety of means to cut down on the actual time required to acquire the spectra: projection NMR, frequency domain sampling, and fast-pulsing NMR. Projection NMR and frequency domain sampling are both techniques that reduce the number of scans needed to acquire a usable spectrum by respectively changing multiple time variables at the same time (rather than just one at a time) or only scanning specific frequencies of interest. Fast-pulsing NMR, on the other hand, reduces the amount of time between successive scans.
With any of these or related methods, the quality of the spectra acquired will be reduced. So, they are based on the question of how much can I get away with and still obtain useful spectra? The answer to that question will depend on what specifically you are trying to accomplish, but none of these methods will be a solve-all panacea.
Another presentation that I found particularly interesting was one by Tim Stevens of the University of Cambridge and a member of the CCPN staff. He discussed a CCPN program currently under development, CcpNmr Clouds, which allows one to determine the structure of a protein by NMR without having to first assign the peaks of its NMR spectra. Anyone who performs NMR knows that assigning a spectrum can be a difficult and time-consuming task (although one that a lot of NMR scientists, including myself, tend to enjoy as an intellectual challenge). The Clouds program is still a work in progress, but the presentation was compelling, as the idea of cutting a huge step out of the structure determination process is very alluring.
After all of the presentations touting the wonders of high speed structure determination, it almost felt as if scientists would be solving structures in their sleep (or maybe people wouldn't be needed at all), but the final presentation--by Sander Nabuurs of the University of Nijmegen in the Netherlands--put all of these advances into perspective. He discussed a study of the quality of a random selection of NMR structures in which it was found that most structures landed in the lower ranges of acceptability. He showed that just by applying a simple refinement technique (water refinement), the quality of these structures improved considerably. Using some examples of particularly bad structures, his presentation explored what parameters NMR structures tend to do worse in and what methods researchers can use to make sure their structures are of the highest quality. Nabuurs' presentation demonstrated the need for even more hands-on human involvement and analysis in the structure determination process, not less, despite the promise of hands-off high speed structure determination methods.
When the information presented from these various presentations is considered together, it appears that high speed structure determination shows a great deal of promise, and should be incorporated into the work of protein NMR scientists. However, at this point at least, these techniques can only be used to produce preliminary structures. Much useful information can be gained from these structures in much less time than before, but the determination of high quality structures will require the incorporation of traditional methods and will still demand many intense hours of hands-on work.
Yep... that about wraps it up....
OK, who am I kidding? Everyone knows the only reason scientists go conferences is for the fun stuff, and CCPN was no exception. The CCPN conference brought us to the surprisingly picturesque Lake District, known for its... well... lakes, but better yet its hills, where one can go on nice and strenuous hikes overlooking the beautiful green English countryside.
Of course, our hike didn't go quite as planned. Although it wasn't so bad that the driving rain and low-hanging clouds obscured our views of the nice scenery, it was a little more worrying that we also lost sight of nearby landmarks and really didn't have any idea where the hell we were or where we were supposed to go. And, it didn't really help that the guy with the map mysteriously and inconveniently disappeared half way through the hike.
Yes, things were looking pretty rough, I tell you. For a while it even seemed like NMR in the UK was about to take a big hit, losing most of its up and coming scientists. However, we didn't want to miss out on the prospect of a night dedicated to sampling the pubs of Ambleside, so we proceeded with rock-hard determination.
I'll spare you the details, but being the clever and resourceful NMR scientists that we are, we eventually made it back to safety, and duly rewarded ourselves with a nice Saturday night out (only to wake up early and hung-over the next morning for more NMR goodness).
I think powerpoint was made for the hungover.