DAQ Evolves into Informatics

The Next Era of Data Acquisition
Contributed editorial appearing in
Scientific Computing & Instrumentation 17:7, June 2000, pg. 13.

In the mid 1990’s, Professor Fred E. Lytle of Purdue University, regarded as one of the principle pioneers of modern analytical science, declared that the measurement and analysis community is entering its third epoch. The first epoch was one of “how can this quantity be measured?” Thus were born the instrumental techniques including spectroscopy, electrochemistry, mass spectrometry, and gravimetric/thermal analysis. The second epoch was one of “how little of this quantity can be measured?” Miniscule sample levels necessitated isolation from interfering species and spurred the development of separation science and the genesis of gas, liquid, and super-critical fluid chromatographies. The recently christened third epoch is one of “how can this quantity be measured where it is?” The strategy employed involves measurement of the interferences concurrent with the desired quantity. Modeling and numerical analysis of the interferences will yield the desired isolated quantity in software. The hardware step of separation (often the most difficult) would no longer be required. Such is the Grail of the third epoch; however, we have an arduous quest ahead of us.

An early by-product of the third epoch is the unbridled proliferation of data acquisition technology capable of measuring myriad signals rapidly. Protocols for acquiring, transmitting, storing, searching, and retrieving data are competing for acceptance. Our current laboratory and enterprise-wide information infrastructures are straining under the shear volume of acquired data. The term Terabyte is entering the layman lexicon. Emergence of the Laboratory Information Management System (LIMS) facilitated the orderly acquisition of data from disparate instruments in remote locations, but this third epoch is revealing the need for even higher-order data acquisition systems.

Analogous to the elemental composition, structure, configuration, and function levels of biologically active molecules, data acquisition at the elemental composition level is a description of type – for instance, unsigned short integer or double precision real. The structure level is likened to the data record profile, i.e., sine wave, single-exponential decay, or array of peaks. A LIMS defines the configuration level of data acquisition. The number, type, and structure of data streams serviced by the LIMS differentiate the configuration of one data acquisition system from another.

Measuring the interferences revealed that the effects of interfering quantities should not merely be measured and removed, but that having knowledge of the interfering quantities themselves is of great importance. This holistic, big-picture view defines the tertiary or high-level function of the system under study. The same values that originally crippled our ability to analyze a seemingly haphazard collection of independent variables are now recognized as essential members of a system comprising a host of interdependent variables. What are needed now are not additional variable isolation methods but the development of variable-integration techniques.

Data acquisition at the LIMS-level distills a collection of related data streams into a structured set of interdependent information and provides a convenient method for information storage and retrieval. Beyond acquisition and cataloging, the information must be analyzed to be useful. Coined in the late 1960’s, the term “informatics” identifies the scientific discipline concerned with the transformation of information into knowledge. Often masquerading under its many aliases, including “information technology” or “information and knowledge management”, informatics has found a renaissance in the health care industries. Pharmaceutical laboratories have long identified that the health effects of a drug are quantified as a set of related side-effects that may be unexpected or even eclipse the magnitude of the targeted effect.

The analysis and management of this valuable information is the subject of “bioinformatics”. Physicians and hospitals that maintain databases containing patient histories and diagnostic test results practice “medical informatics” when seeking to identify the specific cause of a malady. The discipline of “neuroinformatics” is the study of brain function requiring the integration of information from the level of gene expression to the level of behavior. Along with informatics development leaders in industry and academia, the Information Technology Laboratory (ITL) of the National Institute of Standards and Technology (NIST) is developing tools and resources for the transformation of information into knowledge. Modern laboratories and production floors are benefitting from research and development in the areas of Networking Technology, Software Diagnostics, Computer Security, Information Access Technology, and Distributed Computing. Advancements in these divisions are providing the infrastructure necessary to acquire and store large amounts of related data.

The data analysis branch of the ITL contains the Statistical Engineering Division (SED) and the Mathematical and Computational Sciences Division (MCSD). The SED develops statistical design, analysis, and process control procedures while the MCSD focuses its efforts on modern analysis methods used to model large systems of interdependent variables. Recent advances in image analysis and multi-variate modeling are finding broad utility in the laboratory. As the discipline of informatics matures, its developments will extend the capabilities of the LIMS beyond the configuration level into a high-level system that produces usable knowledge. At some point in this third epoch perhaps LIMS will be the acronym for Laboratory “Informatics” Management System.