Automated Interpretation and Results Communication
Developing NMR approaches to provide useful data quickly is only one step towards developing a protocol capable of meeting the high throughput challenge. A complete package must provide data rapidly; automate data processing and analysis; and communicate NMR results in a user-friendly manner. This point was clearly (and somewhat brutally) realized during our early work to automate diffusion NMR binding studies when we were greeted with an imposing stack of data from our first overnight run. Clearly automated processing, interpretation and communication of results would have to figure into any high throughput NMR profile method.
There are a variety of software packages available for NMR aimed at data processing, interpretation assistance and data management. Furthermore, automated data acquisition and processing is available on most, if not all spectrometers via their built-in instrument command languages. As useful as these tools are, they offer little in the way of automated interpretation which is emerging as the single biggest barrier to high throughput NMR. Therefore, many opportunities exist for tighter integration of automated acquisition, processing and interpretation. We have begun to develop automated NMR profile methods using the available tools with the goal of being able to answer simple questions such as: do these compounds bind or did I make the compound I wanted? We have defined a generic approach that starts with a profile method, automates it with the built-in instrument capabilities and utilizes software tools for the interpretation and communication of results. A 'proof of concept' automated binding assay based on diffusion NMR was developed where all data acquisition, processing, interpretation and communication are computer controlled. Figure 4 shows a logical overview of the process with the flow of information from sample wells (left) to HTML documents (right).
Figure 4
The interpreted results (right) are available for browsing as a table where, in the standard manner, a green well means components interacted, a yellow well means an unambiguous determination was not possible and a red well means no interaction detected. This representation can be quickly scanned for positive wells and supporting information (spectra, Kd estimates and specifics on automated interpretation) is available via a hypertext link from each well.
 
©Copyright 2000-2008 Novatia LLC
Disclaimer and terms of usage
e-mail: info@enovatia.com
|
