The ARIA approach is a tool to monitor the NMR structure determination as well as the automatic assignment of the nuclear Overhauser effects (NOE). Only structure generation requires an external structure calculation engine (CNS). Data pre-processing, peak assignment, calibration etc. are entirely performed in Python. This facilitates easier and much faster extensibility of the protocol (e.g. calibration, violation analysis, partial assignment, merging). By using ARIA's Python API, application programmers can integrate new modules or adjust the ARIA core protocols. The ARIA's Python API is described in the folder ARIA for developers.
The generation of conformers is performed by launching a CNS calculation for each conformer. It is possible to run each conformer calculation on a different processor. ARIA and CNS can be run on the same workstation. But another system setup fully compatible with running ARIA is a PC cluster. CNS needs then to run on each cluster processor, but the installation of python, Tix and NumPy is only required on the cluster processor or on the workstation from which ARIA is launched. If ARIA is launched from a workstation external to the cluster, the existence of NIS between the computers helps to run smoothly the calculation.
In order to streamline and simplify the setup of a new project, ARIA provides a standalone graphical user interface to display and edit a project. This was a replacement for the previous, CGI based interface used in ARIA 1.2.
Generally, ARIA handles and represent information such as input data as explicit as possible, thereby avoiding implicit or potentially ambiguous rules. For instance, pseudo atoms are always "decoded" into atom groups, the type of shift assignments is stated explicitly (stereo-specific, floating etc.).
In software engeneering, errors are often also introduced by erroneous or inconsistent input data. This can partially be avoided by forcing the input data into a unified and rigid data-model. We chose the XML standard which is easy to interpret by humans and can be manually corrected if necessary. The demand for input data that accord with the DTD specifications is also to guarantee consistency and cleanliness. The specifications in the DTDs are strict enough to detect oddities or errors during conversion or when parsing the XML files at the beginning of an ARIA run and not eventually during run-time. This reduces occurences of strange program behavior immensely and make tedious searches for reasons of program failure obsolete.
The eXtended Markup Language (XML) is about to become the industry standard for a text-based description of generic data. It facilitates the standardized, portable and human readable storage of arbitrary data. XML files can be edited by using either an XML or a text editor. ARIA uses this format to describe and to store most of the crucial input files. XML is a tag based and allows for validation of documents. Each XML document has to follow its Document Type Definition (DTD). ARIA specific DTDs are found in the "src/xml" directory of the ARIA distribution. The easiest way to create ARIA XML files is by using the conversion facility (see help on conversion in folder Using ARIA). Besides project definition and data-conversion files, ARIA supports XML document definitions for the following data-types: Molecule definition, Spectra, Chemical shifts.
ARIA can read input data through its own format setup, but can also import project from CCPN. For liquid-state NMR data, these two modes of input reading will be conserved in the near future.
CCPN (Collaborative Computing Project for NMR) is a collaborative initiative which provides service to NMR spectroscopists analogous to CCP4 in the X-ray community. The work on the CCPN data model standard in particular, aims at the development of a data model and a common data exchange format enabling the exchange of experimental NMR data between different NMR processing and analysis software packages. Starting from version 2.1, ARIA suports the complete import and export of data and analyses via CCPN
For info materials, publications and further informations, please visit the CCPN Homepage at http://www.ccpn.ac.uk/