Science's Rich Review Literature Charts The Evolution Of Disciplines

Of Disciplines Author: Ricki Lewis Sometimes in science, the best way to move forward is to take a long look backward. For discoveries and data to be put into context and used as springboards for future inquiry, investigators must periodically synthesize and reflect on what has been accomplished, and determine what remains to be learned. A review article is one way to chart the evolution of a scientific discipline. Science has a rich review literature. Of 3,383 journals surveyed in the 1994 Science Citation Index, published by the Philadelphia-based Institute for Scientific Information (ISI), 16,877 of 677,601 authored articles were reviews, accounting for about 2.5 percent of the total. Review articles introduce scientists-in-training to their fields, guide researchers in identifying the work ahead, and give veteran scientists the satisfaction of seeing how their own contributions fit into the big picture. All leading indexing and abstracting services, such as Philadelphia-based Biosis, aid researchers by identifying review articles. ISI also publishes the semiannual Index to Scientific Reviews. "A good review gives a sense of what is known in the field, presented in a clear way, and what you would like to find out in the future," says Alexander Glazer, a professor of biochemistry and molecular biology at the University of California, Berkeley. Glazer, who studies photosynthetic pigments and proteins, won the 1991 National Academy of Sciences (NAS) Award for Excellence in Scientific Reviewing. The NAS award recognizes individuals who have published many outstanding reviews. The annual award is sponsored by Annual Reviews Inc. of Palo Alto, Calif., and ISI. It honors Annual Reviews founder J. Murray Luck, who died in 1993. Annual Reviews is a nonprofit organization that publishes 26 volumes of review articles in a variety of fields. Scientists use review articles in several ways. "I always look for a review article to bring me up to date in an area outside of my specific field," notes James Murray, a professor of animal science at the University of California, Davis. "I also expect a review article to give me a good bibliography, so that if I want more specific information, I have a start on the literature. Finally, I look for a review to let me know the current state of knowledge in a field, and where the major directions will be over the next few years." A scientific review article is often more than a timeline of experimentation. "A well-done critical review influences the direction of research. Annual Reviews can be looked at as an intellectual history of the development of the biological and chemical sciences," says Sam Gubins, president and editor-in-chief of Annual Reviews Inc. RECOGNIZED REVIEWS: Columbia University's Eric Kandel won the NAS award for his neuroscience review articles, which bridge the fields of biology and psychology. Reviews written by Eric Kandel, University Professor at Columbia University College of Physicians and Surgeons and a Howard Hughes Medical Institute senior investigator, for example, are widely credited with bridging biology and psychology, opening up the field of neuroscience. His work explains how to study behavior and learning at the cellular level in the sea snail Aplysia (E.R. Kandel, W.A. Spencer, Physiological Reviews, 48:65-134, 1968). Kandel won the 1988 NAS award for scientific reviewing. TRACKING A FIELD: NAS award winner Ira Herskowitz of UC-San Francisco notes the progress in bacteriophage lambda studies between his first and second reviews. Another NAS award winner, Ira Herskowitz, a professor of biochemistry and biophysics at the University of California, San Francisco, points out that reviews can also introduce new ideas. "In one review article [I. Herskowitz, Nature, 329:219-22, 1987], I proposed a new strategy to inactivate genes. I reviewed existing strategies, and did a comparison to mine," he explains. Herskowitz studies bacteriophage lambda and yeast. A review article may be fewer than 10 pages or more than 100; summarize an entire field or just one aspect of it; appear as one of several reviews in a journal devoted to reviews, or be the sole such effort in a general journal. Reviews mix the comprehensive coverage of an encyclopedia article with experimental results that lend a sense of currency, like that of a research report. THINKING CRITICALLY: Alexander Glazer of UC-Berkeley, an NAS award winner, explains that reviewers in controversial fields must "determine what is ambiguous." The writing process begins with organizing the field's literature. To do this, Glazer piles journal articles on his living room floor, sorted by topic into stacks labeled "A" through "Z." He leaves only a path to the door. Herskowitz also uses the low-tech "piles of papers" method. But Christopher Jacobsen, an associate professor of physics at the State University of New York, Stony Brook, used a computer database of references in writing a recent article because of its length. The article, on soft X-ray microscopes and their biological applications, was nearly 100 pages (J. Kirz, C. Jacobsen, M. Howell, Quarterly Review of Biophysics, 28:33-130, 1995). "The real challenge of writing a review paper is to have all the data and papers at your fingertips, and be able to summarize them," says Carmen Sapienza, a professor of pathology at the Fels Institute for Cancer Research and Molecular Biology at Temple University School of Medicine in Philadelphia. One of his recent papers addressed gene imprinting, the differential expression of a gene depending upon which parent transmits it (C. Sapienza, Developmental Genetics, 17:185-7, 1995). Sapienza, author of eight review articles, finds that if he can't explain a paper in five minutes, or summarize it in a paragraph, then he is not ready to write the review. In addition to being active researchers, review authors must be good writers, using "an accessible style, without complex jargon and without too much specific knowledge of the subject area," according to Gavin Swanson, editor of Trends in Neurosciences (TINS), published by ier Science Inc. in Cambridge, England. TINS is one of 16 "Trends" journals published by ier in different disciplines. Each typically includes several review articles, plus book reviews, news, commentary, correspondence, and brief meeting reports. Herskowitz also emphasizes the importance of clarity in writing. "My credo is, I make no attempt to be comprehensive, but a considerable attempt to be comprehensible," he jokes. He also stresses establishing a paper's focus at the start. A review article often affords the space to include history or folklore. For an article on a young technology, such as transgenic dairy herds, the history might go back only a decade. For topics in physiology, the history may stretch back considerably further. A recent review on liver regeneration, for example, opens with the Greek myth of Prometheus, who was punished for giving humans fire by having his liver pecked out daily by a bird-only to have the organ regenerate overnight (C.J. Steer, FASEB Journal, 9:1396-1400, 1995). This engaging article, by Clifford Steer, a professor of medicine and cell biology at the University of Minnesota Medical School, introduces a series of reviews on liver regeneration. Steer's mythological reference also illustrates the stylistic flexibility review article authors often enjoy. "The latitude depends upon the journal, and how well-written the article is. If a writer takes liberties, but it sounds good, the editor can let the writer get away with it," says Sapienza. His review of genome imprinting includes a conversation with a graduate student to illustrate frequent misunderstanding of this phenomenon, which seems to contradict Mendel's laws of inheritance. Review articles conclude with citation lists, which authors compile from the references used to write the paper. Despite the availability of databases of references, many scientists and graduate students still look to a review article's reference list to get a handy glimpse of a field's evolution. A comprehensive list of citations is an important part of an Annual Reviews article, according to Gubins. The intellectual challenge of writing a review article varies with the topic. "It is easier to write in a field where there is consensus. If the field is controversial, a reviewer must read every article very critically, and determine what is ambiguous and questionable, because a review gives a certain stamp of authority," explains Glazer. An example of a noncontroversial field, according to Glazer, is one in which the evidence measures physical properties, such as structure determinations or syntheses of compounds, assuming that the methods used are well established. He mentions epidemiology as a controversial field. "You have to write about confounding variables, numbers of cases considered, how well a study was conducted, and other factors. Writing a review in such a field is extremely hard work," he adds. Two recently published reviews illustrate Glazer's point. "The nematode Caenorhabditis elegans and its genome" (J. Hodgkin et al., Science, 270:410-4, 1995) is straightforward. Rich in history, the article covers two decades of the inarguable gene-by-gene analysis of this simple animal. In contrast to summarizing gene discoveries is a synthesis of several studies on the effect of dietary soy protein on serum lipid levels (J.W. Anderson et al., New England Journal of Medicine, 333:276-82, 1995). The investigators searched many studies to select those that had rigorous experimental designs. Genome imprinting is another area that is difficult to review, because it is in its infancy, and researchers do not yet know the mechanisms of differential gene expression. "This is a field where there are a lot of hypotheses, but not many answers," says Sapienza. "The reason for most review articles is that there are so many papers, people lose track. With genome imprinting, it is less the case that people want to know where the data are, than they want to know what the data mean." In fact, Sapienza's review emphasizes the field's uncertainty. "I try to alert people that regardless of the number of times they read in Nature or in Science that we've figured out what imprinting is, if you look at the data, none of the hypotheses fit well. It's okay to sit back and be skeptical," he notes. Journal editors identify potential review authors in several ways. "The editors of the 'Trends' journals read the literature, visit laboratories, attend conferences, correspond with the members of their journals' advisory editorial boards, and consult with a network of informal contacts" to find authors, says Swanson. Sometimes researchers approach a journal about writing a review article. Barney S. Graham and Peter F. Wright, professors of medicine at the Vanderbilt University School of Medicine, posed an article idea to editors at the New England Journal of Medicine. "We felt it was a good time to write an overview of where clinical testing of AIDS vaccines was. So we suggested that we be invited," Graham recounts. The article was published in November 1995 (B.S. Graham, P.F. Wright, N. Engl. J. Med., 333:1331-9, 1995). Review articles mark progress. Herskowitz compares the first and second reviews he wrote on lambda (I. Herskowitz, Annual Review of Genetics, 7:289-323, 1973; I. Herskowitz, D. Hagen, Annu. Rev. Gen., 14:399-445, 1980). The first paper identified genes; the second explained how the phage "decides" whether to burst a host cell (lysis), or to integrate harmlessly into its DNA (lysogeny). "The first paper included some speculation about the interaction of two key genes in the lysis/lysogeny decision. It was disproved by the time the second paper was published," Herskowitz says. MANY YEARS IN REVIEW: William Kaufmann of Annual Reviews Inc., which publishes 26 volumes of review articles, poses by a portrait of company founder J. Murray Luck, namesake of the NAS Award for Excellence in Scientific Reviewing. When research reports in a field accumulate rapidly, the time is ripe for a review article. And when review articles in that field begin to accumulate, the time may be ripe for a new Annual Review. Certain signs signal the birth of a new field worthy of its own Annual Review. "There has to be a sense of the field having an identity, a significant body of literature, and a critical mass of people calling themselves researchers in that field," explains William Kaufmann, editor-in-chief, emeritus, of Annual Reviews. Some scientists claim to be too busy doing science to take the time to reflect on it and write a review article. But the effort can be very rewarding, say those who have done it. Notes Scott Tebbutt, research fellow in the department of biochemistry at the University of Otago in Dunedin, New Zealand, who recently published a review article on cystic fibrosis (S. Tebbutt, Molecular Medicine Today, 1:336-42, 1995): "It takes time, but it's a great opportunity to learn as well as make a bit of an impression." Ricki Lewis, a writer based in Scotia, N.Y., is the author of several biology textbooks. She is online at 76715.3517@compuserve.com

Science's Rich Review Literature Charts The Evolution Of Disciplines

Interested in reading more?

Become a Member of