Copyright 2003 Brian E. Coggins and Pei Zhou.
Using the Program
PACES was intended to be easy to use, and we have strived to keep it so throughout the program's development. The following diagram illustrates the flow of information in PACES:
Sequential assignment is carried out as follows:
1. Data Input
The first task is to get resonance data into the program. PACES does not read spectra, but rather takes a list of spin systems with their chemical shifts and assigns those to a protein sequence. This means that picking peaks and editing peaklists must occur outside of PACES, and somehow peaks must be correlated to each other to form spin systems. Ultimately, one ends up with a data file that looks like this:
The reference numbers can be set arbitrarily, but it is recommended that they correspond to the peak numbers in one of the peaklists used for the project. It is not necessary that all of the chemical shifts be supplied (as in this example, which has no HA data; you can also run PACES with individual resonance values missing, scattered throughout the table). Naturally, better results are possible with more data. The User's Manual and the J. Biomol. NMR article contain detailed guidelines on how much data is needed to carry out an assignment.
For users of XEASY, we have automated the spin system assembly process, meaning that you can load in peaklists from different experiments one at a time, and PACES will automatically build up the table above. We are working on adding support for NMRview and Sparky peak lists, hopefully in the very near future. In the meantime, users of these programs will have to build the table some other way, either by hand or using an external program (e.g. Perl). The table may be prepared with Perl in a simple text file format (described in the manual) and imported automatically.
Once the data table has been built, the program is invoked by filling in the blanks in the following dialog box and clicking "Run":
Run times vary depending upon the CPU speed, but average at around 30 seconds for a 100-250 residue protein using a 1.5 GHz processor. A portion of the results page for malate synthase G is shown below:
Residues are listed across the top, and the spin system numbers of possible assignments appear below. In this case, PACES has identified a group of spin systems beginning with number 3 and extending through number 9 as likely assignments for Q3 to R9 (for this example, spin systems were numbered according to their published assignments, making it easier to identify that PACES has located the correct assignments for these residues). A handful of alternative assignments are also possible, such as the one listed at top, with spin systems 381 to 383 assignable at S2 to T4. Due to the nature of the algorithm, the longest such connected segments are generally correct. In this example, the fragment highlighted in yellow could be identified as the most likely based upon its length in comparison to the other possible assignments.
3. Interactive Assignment
Once PACES output is available for a protein, you must narrow down the possible assignments to a single set of definite assignments. This is done using a set of interactive tools which allow you to view information about the spin systems suggested at each position, and which help to connect segments and fill in gaps. When you have chosen a particular spin system for a particular residue from among the possibilities, you can invoke a command to assign the spin system. PACES will then enforce that assignment in subsequent runs, allowing you to reach a single, self-consistent set of assignments.
To help with this process, PACES can prepare XEASY peaklists with its current assignment set and with remaining unassigned spin systems. These can be loaded into XEASY on a UNIX or Linux system along with the spectra to aid with resolving ambiguities or other problems in the data. We hope to add support for NMRview and Sparky peaklists soon.
Some examples of the interactive tools are shown below:
4. Exporting the Results
When you have completed making assignments, PACES can export the data in several formats, including the BMRB NMR-STAR format for chemical shift deposition, a simple text file format convenient for other processing, the TALOS input format, and also as a formatted table for printing. PACES can also prepare preconfigured XEASY peak lists using the assignments; for example, it can create a base peak list for side chain resonance assignment.
Copyright 2003 Brian E. Coggins and Pei Zhou.