Requirements for data recored

English

Requirement for data

An expected output from the operation of samples with analytical instrument is set of result by electrical form.  Most case, result data is saved on the computer as file otherwise directory store into the database system.  Let's call 'data file' for the output recored from the data system.

Data file definition

A data file is corresponding to a sample and is contains raw data file and processed results. Definition of the "raw" data means a series of data stream from instruments during the acquisition. Data streams are containing spectra, voltages, UV/Vis absorption, temperature, solvent composition, pressure and so forth and is usually a function of time.  Data file maybe stored as file on operating system, but also it can be stored in a part of large database management system.

Requirements for raw data

Raw data is the primary data from the data acquisition system and it's meta-data.  Meta-data must contains sample identification, method identification, instrument operation parameters, actual monitoring parameters and meteorologic information.  In addition, detail instrument identification such as serial number, detail configuration information and instrument maintenance information.

For example, acquisition sequence may contains most high level sample information such as sample name, sample id, description of the sample preparation, lab notebook page number, analytical method name, sample size and so forth.  Even a data file is corresponding on single sample, but whole sequence data store to data file is good idea in order to see overall analysis as a history of sample analysis.  Instrument calibration information is always essential for scientific discussions because it is indicating accuracy of the data.

Most of analytical instruments are monitoring voltage as a primary input signal, and digitizing value using ADC converter.  Modern instruments are able to calibrate ADC by using internal reference voltage hocked up on the instruments.  Recording such standard voltage information with raw data will improve further diagnostic for instrument manufacturer.  From the analytical instrument manufacturer's view point, original ADC binary signal combined with data scaling parameter is preferred rather than saved as converted to user space values because more information for diagnostic instrument at the time it was acquired.

Thus raw data should contains:

  1. Instrument identification (such as serial number)
  2. Instrument configuration (such as what sample tray, injector valves, detectors, ion source etc)
  3. Software configuration (such as version number, configured software)
  4. Operator and computer identification
  5. Instrument diagnostic information (such as power monitor, detector running time, source light energy level, reference signal)
  6. Meteorologic data (such as calibration result, calibration sample and operator with approval)
  7. Time stamp since last instrument power on (or reset firmware)
  8. Event in/out data on the series of data.  This is important to synchronize timing information with collaborated instrument such as autosampler triggered by chromatographic signal.

Requirements for processed data

Processed data is a set of spectra, chromatograms, molecular structure or quantity generated by using data system algorithms applied for raw data.  Processed data can be altered by user using data processing user interface by interactively or by batch processing.  Under the regulated environment, each thread of processed data can be altered or stored in the series of history with audit trails.  FDA's 21 CFR Part 11 does not prohibit alter previous data but activity for create/modify/delete need to be recorded with reason and signature.

Processed data should contains meta-data inherit from raw data in order to identify instrument, instrument configuration, software used for acquisition, acquired operator and computer and meteorologic data.  In addition, process operator, computer for data processing and software versions should be recorded.

User login