To answer the questions about storage of electronic records, there are two common ways to approach this (apart from deleting them, of course). The first is the BAD (buy another disk) approach and keep the records on line for as long as you need them — or need another disk. The second is to archive them and keep them near line or off-line. Regulations offer little or no help — they just say how long you have to keep them. So what are we going to do?  

Some help might be at hand, with the recent publication of another regulatory guidance document in the Good Laboratory Practice (GLP) arena that gives guidance to workers in nonclinical studies and bioanalytical. It comes with the snappy title of Establishment and Control of Archives that Operate in Compliance with the Principles of GLP and was published by the Organization for Economic Cooperation and Development (OECD) in June 2007 (1). This is the organization that is responsible for generating GLP regulations outside of the US and Japan. So let's see what this document can do for us in the spectrometry laboratory. 

. . . there were GLP regulations. Archiving of records and study materials generated during the course of a nonclinical health-safety study is a specific requirement of GLP. The rationale for the archive is that maintenance of the raw data associated with a specific study and the specimens generated from that study are the only means that can be used to reconstruct the study and enable the information presented in the final report to be verified.  

The archive is a secure facility or room with an archivist specifically responsible for logging material in and out of the archive. Security and access are major issues within a GLP archive: only authorized personnel have access. This is not the case with GMP regulations. 

The key definitions in this document are as follows: 

Archive: A designated area or facility (such as a cabinet, room, building, or computerized system) for the secure storage and retention of records and materials. 

Archivist: An individual designated by test facility or test site management to be responsible for the management of the archive (that is, for the operations and procedures for archiving). 

Electronic archives: Facilities and systems provided to maintain electronic records as required by the Principles of GLP. 

Electronic record: All original laboratory records and documentation, including data directly entered into a computer through an instrument interface, which are the results of original observations and activities in a study and which are necessary for the reconstruction and evaluation of the report of that study. 

Metadata: Data that describe the attributes of other data. Most commonly, these are data that describe the structure, data elements, interrelationships, and other characteristics of electronic records.  

Migration: The transfer of electronic records from one format, media, or computerized system to another. 

System owner: The manager, or designee, of the department that is most impacted by, or is the primary user of, the system. 

The guidance looks at archiving both in paper and electronic forms; although from many of the definitions here, you can see that the main thrust is toward electronic archiving, and it is this topic that I want to focus on in this column.  

Ok, so who is involved and what do they do in this archiving process? The guidance document outlines several roles and responsibilities, such as test facility management, study director, quality assurance (QA), information technology (IT), and the archivist. I want to focus on just two for the purposes of this discussion: the archivist and IT. If you work in a GMP-regulated laboratory, you can interpret "archivist" as either QA or the laboratory. 

Archivist: The archivist is the gatekeeper for allowing material into and out of the archive. If looking in a GMP laboratory, it's the staff member (or members) responsible for deciding when to move electronic records off-line from the spectrometer. Inevitably, there must be procedures for controlling and recording this process. Therefore, get the structure and the organization of the records right: remember the old saying, "Garbage In, Garbage Out." Therefore, file-naming conventions or a means of indexing these records is important. This can be achieved in a variety of ways and costs: simply by using a directory structure on a dedicated disk, a database, or an export utility if the spectrometer software has this facility, or using a scientific data management system (SDMS). 

IT Personnel: IT personnel involved in archiving operations (such as ensuring integrity of electronic records) should be trained adequately and should understand GLP (or GMP) as it relates to their work. They must work closely with the archivist and have either SOPs or written service level agreements (SLAs) to cover this work. Establishment of physical security of the server and the room where it is housed and the logical security of who can access the electronic records are vital, as we will see next.  

The guidance discusses the facilities at some length. However, let's focus on the facilities for storage of electronic records.  

The components that provide storage of unique electronic records should be physically secure.  

The computerized archive facility should have processes to prevent unauthorized access and virus protection. This should include both logical and physical security elements. 

The security controls necessary to restrict access to electronic records usually will be different from those applied to other record types. Because many electronic storage media can be reused (or overwritten), measures should be implemented to ensure that records cannot be altered or deleted. 

For electronic archives, as a minimum, deletion or alteration of electronic records in electronic archives should be avoided. This is too tactfully worded in my view; I would interpret this as, where it says "should," read "must." 

In addition, the environmental conditions of the room where the archive servers are housed should be monitored and have an uninterruptible power supply (UPS) to protect the system from power loss. Not mentioned explicitly in the guidance, the computer hardware should be sufficiently resilient to overcome common problems such as disk or processor failures. In addition, disaster recovery including fires and floods should have contingency measures that are documented and work to ensure that electronic records are not lost in the event of failure or disaster. 

The requirements for archiving of electronic records have additional needs compared with paper, which the guidance presents in five main subject areas and I will discuss in the following sections.  

Decision Regarding Electronic Records Retention 

There must be a conscious decision to work with electronic records and also retain them. This can require the need for file naming conventions if the current application software cannot do this. 

The decision to retain electronic records has major ramifications for any laboratory and will influence the choice of storage media. One key issue is the off-line storage media change. Look at the issue with the current DVD replacement wars: BluRay or HD-DVD — which one will you place your money on? From my perspective, the best way to store records is online to avoid the media problems.  

This now brings the next issue: Can we read the data in the same format for the length of the retention period? We might need the original software, or we might need to migrate records to a new file format to ensure readability. Computer technology is developing rapidly, and devices capable of reading storage media in common use today probably will not be available in the future. If records need to be migrated, procedures must be in place to ensure the integrity of data from one format to another. Do not be surprised if data elements are lost during the migration, but this needs to be documented and the impact assessed. This is reflected in the guidance as follows: Electronic media can be discarded when the media itself no longer permits evaluation (due to hardware or software issues) provided the disposal is authorized, documented, and electronic records are migrated and any record losses documented. 

Defined Archive Area on a Computerized System 

Electronic records can be moved from the production part of a computerized system to a discrete, secure archive area on the same computer system (physically separated, such as file record systems), or explicitly marked as archived (logically separated, such as database record systems). Archived records should be locked so that they cannot be altered or deleted without detection. 

Again, in a pragmatic way, the guidance offers the following advice; viewing electronic records without the possibility of alteration or deletion of the archived electronic record or replicating within another computerized system does not constitute retrieval of a record. This offers a major advantage over a paper archive — you must be inside the archive to see the physical paper record. In the electronic world, anyone could have read-only access to the electronic record.  

Complete electronic records must be entered into the archive system. This includes all the data and metadata, including electronic signatures required for reconstruction of a study, otherwise the migration is useless. When electronic records are being transferred into the archive, the person doing this must ensure that there is agreement between what is documented to be transferred and what actually has been transferred. Ideally, the archivist should be the system owner and have an SLA with IT to manage the system. 

Maintenance and Preservation of Electronic Records 

This is the difficult bit — we have about 2000 years experience handling paper records but less than 50 years with computerized systems, less if you consider that it was only in the 1980s when computers became widely available to many people. Procedures should be in place to ensure that essential information remains complete and retrievable throughout the specified retention period, and this also includes the mechanisms for backing up the systems.  

What might be problematic is the following statement: "If the record medium requires processing in order to render the retained records into a readable format, then the continued availability of appropriate equipment should be ensured." This is the problem because how many vendors provide a future proof method of maintaining electronic records? Standards to help migration are emerging but need to be adopted widely — alternatively, we are back to migration. 

If electronic record migration is necessary, the process of migration should be fully documented and validated to ensure complete and accurate migration of the original records before they are lost or destroyed. If it is impossible to migrate the records to new electronic media, it might be necessary to migrate to paper records.  

I have summarized the OECD GLP guidance document on archiving in this column; it is an important guidance because for the first time, it attempts to interpret from a regulatory perspective the basic operating principles for an electronic archive. I have not summarized the whole document but just focused on the electronic record portion. For those readers who need more information, I recommend reading the original document. 

Articles

Focus on Quality

Paper is easy to archive, but what about archiving electronic records? What do you do with all the electronic records that are generated? In this month's installment, columnist Bob McDowall explores the issue of electronic records management and looks at the recent guidance issued by the OECD for GLP laboratories.

Archiving Electronic Records

Fourier transform–infrared (FT-IR)spectroscopy technology has progressed considerably over the past two decades, and it is now a relatively established analytical technique for process monitoring in addition to being a standard tool in the laboratory. The inherent design of FT-IR systems makes them preferable for use as a process monitoring and analysis tool, particularly in the life science industries, which is a promising market. 

The core of any FT-IR instrument is the interferometer, which by its very nature provides its own internal calibration. The ability to eliminate the need for external calibration is obviously a major advantage for process instrumentation. In addition, FT- IR provides much higher signal-to-noise ratios in comparison to dispersive IT, which allows for much more rapid analysis. 

The process life science market for FT-IR includes pharmaceutical, agriculture & food, and organic chemicals industries.  The best known application is probably reaction monitoring, a significant portion of which falls under the definition of process analytical technology (PAT) in the pharmaceutical industry. FT-IR is also useful for identifying the level of key nutritional components in foods. The combined worldwide market for these applications for process FT-IR is estimated to be $20 million, and it is expected to see double-digit growth for some time to come. 

 (MAP) report program.  For more information, contact Stuart Press, Senior Consultant, Strategic Directions International, Inc., 6242 Westchester Parkway, Suite 100, Los Angeles, CA 90045, (310) 641-4982, fax: (310) 641-8851, 

Departments

Fourier transform–infrared (FT-IR) spectroscopy technology has progressed considerably over the past two decades, and it is now a relatively established analytical technique for process monitoring in addition to being a standard tool in the laboratory. The inherent design of FT-IR systems makes them preferable for use as a process monitoring and analysis tool, particularly in the life science industries, which is a promising market.

Market Profile: Life Science Process FT-IR

Science (n): The process of learning about the natural universe by observation andexperiment. 

At least, that's what I teach my students, usually on the first day of class. Science is, ironically, difficult to define precisely, but this is the definition I've developed over the years. If anyone has any suggestions for improvement, let me know. I also teach my students that if theory doesn't agree with nature, there are two choices: change the theory, or change nature. Unfortunately, all attempts to change nature have failed. Our only choice? Change the theory. That, in a nutshell, is why we have quantum mechanics. In this first part of a multipart series, we will review the failures of classical mechanics that necessitated the development of a new theory of nature. 

Isaac Newton first published his three laws of motion in 1687, in 

. Although they took some time to be accepted, ultimately it was realized that the three laws did indeed accurately describe the motion of objects. The predictive nature of the three laws was so good that it inspired Pierre-Simon Laplace, the famed French mathematician and astronomer, to famously remark that if we knew the position and velocity of every bit of matter, we could predict the future of the universe. 

However, as the 19th century progressed, certain experiments produced results that could not be explained in terms of the understanding of nature that science held at that time. Here, we will review the main issues. 

In 1860, Robert Bunsen (of Bunsen burner fame) and Gustav Kirchhoff (of Kirchhoff's laws fame) invented the spectroscope (Figure 1). One of the noteworthy findings, noted almost immediately, was that some samples absorbed only particular colors of light. Bunsen and Kirchhoff quickly concluded that the colors of light absorbed (or emitted, depending upon the temperature of the sample) were specific to the elements in the sample, and less than a year later discovered the elements rubidium and cesium by noting new colors of light not previously associated with any other known element. 

Figure 1: One of the earliest models of spectroscope (1). 

Of course, there is an obvious question — why? Why do different elements absorb or emit only certain colors of light? Also, what determines which colors of light are absorbed or emitted? Unfortunately, nothing in Newton's laws of motion, or law of gravity, or even Maxwell's equations of electromagnetism (published in 1861) could explain why. 

The element hydrogen had a particularly simple spectrum. This was arguably defensible, because it was known even then that hydrogen was the simplest element. However, there was no theoretical explanation for the particular colors that hydrogen did absorb or emit (Figure 2). 

Figure 2: Emission spectrum of the element hydrogen. The leftmost line is in the far violet and sometimes is not detectable by the human eye. A major issue in the late 19th century was exactly why these colors, and no others, are emitted by hydrogen. 

The simplicity of the spectrum did inspire some investigation, and in 1885, Swiss mathematician Johann Jakob Balmer announced that the wavelengths of the lines of light emitted or absorbed by hydrogen could be predicted by the following formula: 

 was 3, 4, 5, and 6 for the lines going from red to violet, respectively. This formula is more well known in its reciprocal form as 

Collectively, these four lines of light are called the Balmer series of hydrogen, in Balmer's honor. Three years later, a Swiss physicist, Johannes Rydberg, generalized Balmer's formula to include light emitted in other regions of the spectrum; his formula is 

 is a constant now known as the Rydberg constant. This formula predicts the wavelengths of all light, in any region of the spectrum, given off by hydrogen. 

Great — we can now predict the spectrum of hydrogen. But the main question still remains: why does this work? What theory of nature can we invoke to explain this? Unfortunately, the understanding of nature at that point in history was insufficient to explain spectra and, no matter how simple, the spectrum of hydrogen. 

Although it had been observed earlier, the first details of a phenomenon called the photoelectric effect were published by Heinrich Hertz in 1887. The photoelectric effect is simple: when light of a certain wavelength was shined on a metal surface, electrons were emitted. Hertz apparently reported his observations but did not pursue the topic, nor did he attempt any explanation of the phenomenon. 

The most detailed observations of the photoelectric effect were made by Philip von Lénárd in 1902. He showed that the energy of the emitted electrons (which he called "cathode rays") was inversely proportional to the wavelength of the light used to promote emission, but independent of the intensity of the light. It also was shown that the intensity of the cathode ray emission was dependent upon the intensity of light used. 

This was all acceptable information, but the fundamental question remains — why? Why did these effects appear? One of the more questionable observations had to do with the intensity of the light used. If light were a wave (see below), then the energy of the wave is related to its intensity. This suggests that the more intense the light source, the greater the energy of the emitted cathode rays (electrons). However, it was demonstrated that intensity was not related to the energy of the emitted electrons; rather, the wavelength of the light used determined the energy of the emitted electrons. This did not make sense in terms of the understanding of nature at the time. Once again, the theories of nature at that time were insufficient to explain a measurable phenomenon. 

The Nature of Light and Blackbody Radiation

The nature of light is a topic that has been covered recently in this column; readers interested in details can refer to earlier installments (2). The issue of whether light is a particle or a wave supposedly was settled in the early 1800s when English polymath Thomas Young used the so-called double-slit experiment to produce interference fringes, evidence that light was a wave. So the matter rested. 

In the mid 1800s, interest in the luminescent properties of blackbodies arose. A blackbody is a perfect absorber of light. Because absorption and emission were recognized as opposite processes, another definition is that a blackbody is a perfect emitter of light as well. Blackbodies emit light when hot, so scientists eventually studied the properties of light emitted from hot blackbodies. 

At first, one might think that the emission of a wave of light of any wavelength is equally probable, so a plot of intensity emitted versus wavelength of light ought to be a straight horizontal line (marred by peaks and valleys due to experimental error and the imperfection of real materials). However, experimental measurements showed a variation in the intensity versus wavelength plot. The actual behavior looked something like that shown in Figure 3. The intensity is zero at very small wavelengths, and then increases to some maximum whose wavelength depends upon temperature, and then decreases asymptotically as the wavelength of light increases. 

Figure 3: Plots of the intensity of light emitted versus wavelength for a hot blackbody. Although the general curvature is similar, the actual intensity pattern depends upon the temperature of the blackbody.  

Why does emission of light from blackbodies follow this behavior? Scientists at the time could not explain the behavior of blackbody radiation (sometimes called cavity radiation) using the theories of the time. There were some observations, however. 

 is not only temperature-dependent, but it follows a simple mathematical relationship: 

This relationship is called Wien's displacement law, after Wilhelm Wien, the German physicist who first annunciated it in 1893. 

The total power radiated per unit area of blackbody, 

 is related to the area under the blackbody intensity curve. It was found experimentally by Jožef Stefan and thermodynamically by Ludwig Boltzmann that the total power radiated per unit area is proportional only to the fourth power of the temperature: 

 is called the Stefan–Boltzmann constant, and its value is 5.671 × 10

These relationships were verifiable experimentally, but the question remains: why do these relationships occur? There were some attempts to derive theoretical explanations for the behavior of blackbody radiation. Wien (of Wien displacement law fame) derived an expression in 1896 that became known as Wien's distribution law (not to be confused with his displacement law). He derived the following expression: 

 are constants to fit the experimental data. Wien's distribution law was only partially successful, agreeing with experiment on the high-frequency (low-wavelength) end of the blackbody intensity curve. 

A different attempt was made by English physicist John William Strutt, 3rd Baron Rayleigh. In 1900, he derived a different expression that was refined in 1905 by fellow British physicist James Jeans. Their function is 

 is Boltzmann's constant and λ is the wavelength of the light emitted. This function, known as the Rayleigh–Jeans law, also was imperfect: it agreed well on the high-wavelength (low-frequency) side of the experimental blackbody radiation distribution curve. 

So what we have at this point is multiple attempts but no satisfactory theoretical understanding of why blackbody radiation behaves the way it does. The theories of nature at that time were insufficient to explain this phenomenon. 

In the late 1800s, Dutch physicist Heike Kamerlingh Onnes pioneered the study of materials at temperatures close to absolute zero. (He was, in fact, the first person to liquefy helium, in 1908.) Upon determining the heat capacity 

 of various materials at various temperatures, it was found that the specific heat capacities approached zero as the temperature approached zero. Furthermore, as one slowly increased the temperature, the heat capacity increased and was proportional to the third power of the absolute temperature: 

 approach zero must be the case, thanks to the third law of thermodynamics. But why the third-power dependence upon absolute temperature? Scientists couldn't explain why this must be so. 

The reason that these phenomena could not be explained was because the physical theories of nature at that time were incomplete. For the first time, scientists were probing matter at the atomic level. What we realize now is that atoms behave slightly differently than bulk matter, and so different physical theories are needed to explain them. (Rest assured that today's science does have adequate theoretical explanations for all of the phenomena presented here!) 

To explain these phenomena, scientists needed to understand two things: first, that light acts like a particle in some of its properties; and second, that electrons and other tiny particles of matter act like waves in some of their properties. Until the false dichotomy of "particle or wave" was removed, science couldn't explain these phenomena. Science needed quantum theory and quantum mechanics. Ultimately, if you do spectroscopy, you're doing quantum mechanics. 

In the next installment, we will see how these new theories, and the insights that come with them, apply to the phenomena discussed here. 

A similar but more detailed version of this story is given in my textbook 

 (3). It is my opinion that most textbooks in physical chemistry don't devote sufficient space to the failings of classical mechanics that required the development of quantum mechanics — that was one reason I wrote my own book! 

 is a professor of chemistry at Cleveland State University in Ohio. Many of his "Baseline" columns have been reprinted in book form by SPIE Press as 

, available through the SPIE Web Bookstore at 

(published in May 2006), is available from SPIEPress. He can be reached at 

The Baseline

In this first part of a multipart series, columnist David Ball reviews the failures of classical mechanics that necessitated the development of a new theory of nature.

Spectroscopy Is Applied Quantum Mechanics, Part I: The Need for Quantum Mechanics

After all the tales of "thunder lizards" roaming the vast landscapes millions of years ago—enormous predators with giant razor-sharp teeth and voracious appetites—could they have been little more than big chickens? 

Protein links have been found between tyrranosaurs and modern chickens. 

It has long been agreed by paleontologists that modern birds are the descendants of dinosaurs, specifically theropods. Theropods (meaning "beast feet") first appear during the Carnian age of the Late Triassic nearly 220 million years ago, with tyrannosaurs falling under that classification. The first appearance of a birdlike creature in the era of dinosaurs is the Archaeopteryx, which lived in the late Jurassic Period around 155–150 million years ago. The Archaeopteryx possessed feathers and wings like a bird, but it also had teeth and a skeleton similar to a small carnivorous dinosaur, giving it both bird and theropod dinosaur features. Many scientists believe that because all descendants of a single common ancestor are related, and birds have evolved from ancient dinosaurs and are technically dinosaurs themselves, then dinosaurs aren't really extinct. 

Using mass spectrometry, researchers have found protein links between a 68-million-year-old 

 fossil and the common chicken. Scientists found that this particularly well-preserved 

 fossil contained proteins of the animal, which they were able to recover. This discovery came to light in 2003 when Mary Higby Schweitzer from the University of North Carolina recovered soft tissue from a 

 bone found in Montana. With those findings, John M. Asara of Beth Israel Deaconess Medical Center and his research team in Boston were able to analyze the collagen found in that bone, matching it closely with collagen found in modern chicken bones. The major breakthrough here is that, while it has long been thought that birds evolved from dinosaurs based upon the architecture of the bones, now there is more solid evidence that they are related because of the similarities in their sequences. 

Asara's team used an ion trap mass spectrometer, which can determine sequences derived from ancient organisms from peptide fragmentation patterns. After the peptides were measured for mass, Asara and his team isolated and fragmented them to uncover seven separate strings of amino acid.  Asara then compared the amino acid sequences to living species, three of which matched chickens, while two others matched a newt and a frog. 

These new findings are a breakthrough on two fronts. One, it was never before thought possible that protein sequences would be found in a fossil that is over one million years old.  Two, it helps prove conclusively that birds and dinosaurs are indeed related. It's something to think about the next time you order chicken. 

Spectroscopy editor Brian Johnson discusses recent mass spectrometric findings linking dinosaurs and modern birds.

Spectroscopy-12-01-2007

Archiving Electronic Records

Market Profile: Life Science Process FT-IR

Spectroscopy Is Applied Quantum Mechanics, Part I: The Need for Quantum Mechanics

End of the Spectrum: Spectroscosaurus Rex