Atlas of Protein Sequence and Structure

The Atlas of Protein Sequence and Structure, hereafter the Atlas, was a computerized collection of protein sequences distributed in the form of a paperback book published from 1965 to 1979. Margaret Oakley Dayhoff, a researcher who used computers to study biochemistry, was the primary author. Most editions of the Atlas included chapters of original research and computer programs for analyzing protein-sequence data to determine protein structure and function and derive insights about evolution at the molecular level. Before the establishment of the Atlas, biochemists sometimes published specialized sequence collections for comparative research about a given protein. The Atlas was the first comprehensive collection of all published protein sequences derived from diverse organisms. The Atlas laid the foundation of what is, as of 2026, a leading international protein-sequence database, enabling numerous lines of research, such as predicting the function of unknown proteins, which in turn informs scientific technologies and advances such as gene editing.

The Atlas contains sets of homologous protein sequences, which are sequences of the amino acids in comparable proteins originating from organisms with a shared evolutionary ancestry. Scientists determined the first protein sequences in the 1950s, prompting biochemists to begin comparing molecularly similar sequences to understand the functions they serve in biological organisms. Proteins are large molecules in biological organisms that perform a vast array of functions, such as facilitating the transport of oxygen in red blood cells. Proteins are composed of subunits called amino acids, which are linked together in chains. A human hemoglobin molecule, for example, is composed of four amino-acid chains. The amino acid sequences of homologous hemoglobin proteins are more similar for more closely related species. For example, the hemoglobin sequences in humans and chimpanzees are identical, but between humans and gorillas there is one amino acid difference in each of the four amino-acid chains. Sequences in the Atlas were often arranged in such a way that researchers could compare homologous sequences side by side to make inferences about their structure, function, or evolutionary relationships.

While the Atlas was a collaborative endeavor, involving a full team of co-authors, engineers, and other staff, historians of science credit Dayhoff and Richard V. Eck, a researcher with a background in developing mathematical models for studying cancer, with its initial development and leadership. In 1960, the National Biomedical Research Foundation, or NBRF, a private non-profit institution located just outside of Washington, D.C. that focused on the intersection of technology, biology, and medicine, hired Dayhoff under a National Institutes of Health grant to explore the uses of computers in biology and medicine. Dayhoff and a colleague wrote a computer program to expedite standard laboratory methods for protein sequence determination. At that time, researchers determined protein sequences by chemically breaking full proteins into fragments, determining the sequence of those fragments, and then reassembling them. Dayhoff and Robert S. Ledley, the founder of NBRF, devised a program to automate and expedite the reassembly of fragments into whole sequences. In the same year, NBRF hired Eck, who was concurrently working on a similar computer program for partial sequence reassembly at the National Cancer Institute in Bethesda, Maryland. Upon accepting a position at NBRF, Eck began comparative protein research where he analyzed all the known sequences of homologous hemoglobin variants across species in an attempt to decipher the specific correspondence between deoxyribonucleic acid, or DNA, and ribonucleic, or RNA, sequences and the sequences of amino acids in proteins. DNA is the double-stranded molecule that carries hereditary information for an organism’s development and functioning. RNA is the single-stranded molecule that carries hereditary instructions from DNA to direct protein synthesis. As a result of their work on comparative protein sequences, Eck and Dayhoff started to compile published amino-acid-sequences data.

Before Dayhoff and Eck started publishing the Atlas in paperback format, the on-site sequence collection they created and used internally at NBRF in the early 1960s exhibited many features of a database. Because Dayhoff and Eck’s early work comparing proteins was computational, meaning that it involved using computers to perform calculations or analyses, they stored and tabulated sequence data in ways that a computer could most easily interpret. Dayhoff and Eck stored the collected sequence data on punch cards, which are physical cards perforated in systematic ways to represent data and computer programs, so that they could feed the cards into computers of that time to read. Each card constituted an entry in what they called the Amino Acid Sequence Library, which was the computerized complement of the published print edition they kept for internal use. Dayhoff and Eck also adopted a one-letter notation for amino acids instead of the usual three letter code, which saved computer memory and made the data more readable on compact printers. Thus, their internal punch-card library exhibited many features of a database.

In 1962, Linus Pauling and Émile Zuckerkandl, a pair of researchers studying biochemistry who were unaffiliated with NBRF, proposed using differences in homologous protein sequences to measure evolution at the molecular level, which Dayhoff and Eck later cited in the preface of the first edition of the Atlas as a significant motivation for its publication. Up until the 1950s scientists primarily studied homologous proteins by examining their physical shapes. Pauling and Zuckerkandl posited with their 1962 molecular-clock hypothesis, however, that over the course of many thousands of years, protein sequences accumulate mutations at an approximately constant rate. They further proposed that researchers could use the number of differences between two homologous protein sequences to measure the evolutionary distance, or the number of years since the most recent common ancestor, between them. In a paper titled “Amino-Acid Composition of the Polypeptide Chains of Gorilla Hemoglobin,” Zuckerkandl and Walter A. Schroeder, a colleague of Pauling’s, used the number of amino acid differences between human and gorilla hemoglobin sequences to derive the evolutionary distance between the two organisms.

In 1965, Dayhoff, Eck, Marie A. Chang, and Minnie R. Sochard published the first paperback edition of the Atlas. In the introduction, they note that they intended for the Atlas to be a comprehensive reference containing every amino acid sequence for every protein with available data. At just under 100 pages, the Atlas contained sixty-five protein sequences and included classification by protein name, species name, reference to the original source species, and the protein sequence itself for each of the sixty-five entries. The authors printed the protein sequences using both the three-letter amino-acid abbreviation and the custom one-letter system that they had adopted for the punch-card library.

The first edition of the Atlas received both acclaim and criticism, the latter of which historians argue was often based in gender bias. Before the advent of computers, and as early as the eighteenth century, people, and especially women, worked for others to perform repetitive calculations by hand. When Dayhoff first published the Atlas, she solicited feedback from scientists, including Nobel laureates, whose work her team included in the first edition. Several of them responded to Dayhoff by praising the Atlas as a resource. According to Bruno J. Strasser, a historian of science and professor at the Université de Genève (University of Geneva) in Geneva, Switzerland, other scientists undercut similar compliments by referring to Dayhoff and her coauthors, who were primarily female, as compilers who did not do experimental work. Strasser notes that historical gender roles in computing may have influenced the early valuation of the Atlas.

The second edition of the Atlas, released in 1966, not only included sequence data but also introduced novel methods of sequence analysis related to molecular evolution, among other research findings. Dayhoff and colleagues published the second edition only one year after the first edition, yet it doubled in size as it accommodated what Dayhoff referred to as the explosive growth of published sequenced proteins. She and her team arranged protein sequences based on how they would be most useful to researchers focusing on comparative analysis. In the second edition, they used fold-out pages to spread out long, unbroken sequences across a single, extra-wide page. That way, homologous sequences, like hemoglobin sourced from various organisms, were juxtaposed in such a way that users could identify conserved regions across the protein, which hinted at functionally important segments within the protein. Those pages employed the single-letter notation established in the first edition of the Atlas so that readers could more easily discern amino acid differences among the homologous sequences. The second edition also contained a breadth of supplemental information about protein structure, methods for comparative analysis, and other tools for inferring evolutionary relationships.

NBRF published three additional editions of the Atlas between 1968 and 1972, all of which were widely praised, especially by researchers doing comparative research who had previously been trying to keep up with all the known sequences on their own. Book reviewers of the third through fifth editions lauded the Atlas as a mandatory resource for evolutionary analysis of protein sequences, a nascent field of science later called molecular evolution. In reviews of the Atlas, critics such as Edward Cox and Walter Fitch agreed that even though the computational methods in the text were high quality, the Atlas did not contain many alternative approaches, potentially because most co-authors across all five editions were internal to NBRF rather than from several different organizations. In 1970, one book reviewer, Robert Hersh, highlighted that the cost of the Atlas had quadrupled in three years to US $12.50, equivalent to $107.92 in 2026, predicting it would soon become prohibitively expensive for individual investigators or laboratories. That change reflected the challenge Dayhoff faced as she grappled with rising operational costs due to the growing number of sequences.

Even though Dayhoff’s expertise and the quality of the Atlas were well-regarded, per Strasser, her attempts to sustainably scale the protein-sequence collection and tracking with the cooperation of sequence-publishing laboratory researchers proved challenging. With the cost of curating sequences from the published literature on the rise, in parallel with the accelerating growth of published sequences, Dayhoff attempted to establish a reciprocal exchange with biochemists involved in the science of protein sequences. For several years, she wrote letters to those scientists, proposing a data-sharing arrangement in which laboratory researchers would voluntarily share their sequences for free to be included in the Atlas in exchange for a complimentary copy . According to Strasser, one of the possible reasons Dayhoff’s data-sharing arrangement failed was because she violated lab scientists’ sense of control over and recognition for their sequence data. Dayhoff did not just curate and reprint sequence data, an activity that, according to Strasser, researchers widely praised ; she also frequently analyzed submitted data for publishable insights. Those insights were not solely derived from comparative research, which necessitated a collection such as Atlas, but also from her timely analysis of recently published individual sequences. According to Strasser, that behavior caused resentment among laboratory scientists in particular who felt that their research opportunities were being preempted. In a personal correspondence, one researcher even described the Atlas as Dayhoff’s private hunting grounds.

When she failed to establish a data-sharing arrangement, to increase revenue, Dayhoff raised the price of the print editions of the Atlas and started selling the sequence data in magnetic tape formats that could be read directly by a computer. For instance, in 1977, she sold a tape containing the sequence data for $400. That magnetic tape required purchasers to agree not to redistribute the data, which Dayhoff copyrighted. According to Strasser, even though NBRF was a non-profit research institution and the magnetic tape was sold at what he states was a modest price, the fact that she had copyrighted the sequence data, limited its redistribution, sought revenues from it, and used data submitted to the Atlas for her own research violated the laboratory researchers’ ethical standards.

In the publication of the fifth edition and additional supplements, from 1972 onward, Dayhoff and her colleagues organized the Atlas using the concept of protein superfamilies. At the time, scientists had well-established methods for classifying groups of similar proteins that appeared in many different organisms into what they called families. However, in their article “Evolution of Sequences Within Protein Superfamilies,” Dayhoff and her colleagues at NBRF introduced superfamilies, which are higher-level groups of sequences that are all recognizably related by a common ancestor. They calculated that most protein superfamilies have a common ancestor extending throughout all eukaryotes, which are organisms such as plants, fungi, and animals whose cells contain a nucleus, or throughout a major division of bacteria, which are all single celled and do not contain a nucleus. They found that some superfamilies even occur in all organisms. The superfamily classifications served more than organizational purposes; they offered theoretical insights into evolution, revealing the conservative nature of the molecular evolutionary process. Those classifications highlighted how protein structures endure across species divergences.

NBRF published three supplements in the years after the final edition of the Atlas, and in the last supplement, in 1978, Dayhoff established a novel computational tool called the Point Accepted Mutation, or PAM, matrix. The matrix enabled the comparison of protein sequences to allow researchers to infer evolutionary relationships on larger timescales. The PAM matrix improved the accuracy of classification of proteins into superfamilies and allowed Dayhoff and others to meaningfully contribute to phylogenetics, or the organization of related organisms into a hierarchical tree-like structure based on their descent from a common ancestor. The PAM matrix remained the dominant method for assigning scores of similarities between two protein sequences until the early 1990s.

After the last paperbound supplement in 1978, the Atlas was thereafter only available in digital formats, allowing for continuous updates and the development of a new, freely accessible resource. Also in 1978, Dayhoff devised a computer retrieval system that allowed NBRF staff to store and retrieve the existing data from the Atlas but also add and organize new entries as more sequences became available . After obtaining funding to further develop that computer retrieval system into a free online database, NBRF launched that database in 1984 under the name Protein Information Resource, or PIR, which contained the sequences originating from the Atlas collection and many other collections.

The Atlas served as a model for several other biological databases, many of which researchers still use as of 2026. From early on in its history, NBRF maintained the Atlas data in an on-site collection using computer technology. Over the years, the internal analogue database precursor evolved from a set of computer punch cards to a dedicated computer retrieval system. NBRF further developed the computer retrieval system into the PIR, which served as the home of the primary protein sequence database in the United States for thirty years. Just as the Atlas had also been, the PIR was also an integrated resource with many computer programs for finding, analyzing, and processing protein sequences. In 2002, several former staff of NBRF, which was dissolved in 2010 upon founder Robert S. Ledley’s retirement, and several international partners launched UniProt, a single worldwide database of protein sequence and function. As of 2026, UniProt allows scientists all over the world to take an unknown protein sequence, compare it to the thousands of known protein sequences in the worldwide database, and determine the ways in which the unknown one is alike and different from known types. From that data, scientists can quickly and accurately deduce a protein’s evolutionary history and its relationship to various forms of life.

The Atlas opened the door for comparative and quantitative analysis at the molecular level, establishing the foundation for data-driven biological research by serving as a comprehensive, open collection of protein sequence data before such resources were commonplace.

Sources

1. Abbate, Janet. “Women and Gender in the History of Computing.” IEEE Annals of the History of Computing 25 (4) (2003): 4–8.
2. Brewer, Nic. “Protein Information Resource.” Embryo Project Encyclopedia (2025-08-06). ISSN: 1940-5030 https://hdl.handle.net/2286/R.2.N.202529.
3. “Celebrating a Pioneer in Bioinformatics.” Nature Computational Science 5 (3) (2025): 187–187. https://www.nature.com/articles/s43588-025-00784-y (Accessed October 23, 2026).
4. Cox, Edward C. “[Untitled].” Review of Atlas of Protein Sequence and Structure 1967–68, Maragaret Oakley Dayhoff and Richard V. Eck. American Scientist 57 (1969): 36A.
5. Dayhoff, Margaret Oakley. Atlas of Protein Sequence and Structure (Volume 5). Washington, D.C: National Biomedical Research Foundation, 1972.
6. Dayhoff, Margaret Oakley, P. J. McLaughlin, Winona C. Barker, and Lois T. Hunt. “Evolution of Sequences within Protein Superfamilies.” Die Naturwissenschaften 62 (4) (1975): 154–61.
7. Dayhoff, Margaret Oakley, Richard V. Eck, Marie R. Chang, and Minnie R. Sochard. Atlas of Protein Sequence and Structure. Washington, D.C: National Biomedical Research Foundation, 1965. https://ntrs.nasa.gov/api/citations/19660014530/downloads/19660014530.pdf (Accessed October 23, 2026).
8. Dayhoff, Margaret Oakley, Winona C. Barker, R. M. Schwartz, Bruce C. Orcutt, and Lois T. Hunt. “Data Base for Protein Sequences.” AFIPS ’76: Proceedings of the June 7-10, 1976, National Computer Conference and Exposition (1976): 261–66. https://dl.acm.org/doi/abs/10.1145/1499799.1499841 (Accessed October 23, 2026).
9. Eck, Richard V. “A Simplified Strategy for Sequence Analysis of Large Proteins.” Nature 193 (1962): 241–43.
10. Fitch, Walter M. “[Untitled].” Review of Atlas of Protein Sequence and Structure, 1972 (Volume 5). Systematic Biology 22 (2) (1973): 196–196.
11. Hersh, Robert. “[Untitled].” Review of Atlas of Protein Sequence and Structure, 1969 (Volume 4). Systematic Zoology 19 (2) (1970): 192.
12. “History [PIR - Protein Information Resource].” https://proteininformationresource.org/pirwww/about/aboutpir.shtml (Accessed May 26, 2025).
13. Morgan, Gregory J. “Emile Zuckerkandl, Linus Pauling, and the Molecular Evolutionary Clock, 1959–1965.” Journal of the History of Biology 31 (2) (1998): 155–78.
14. Orcutt, Bruce C., David G. George, and Margaret O. Dayhoff. “Protein and Nucleic Acid Sequence Database Systems.” Annual Review of Biophysics and Bioengineering 12 (1) (1983): 419–41.
15. Pearson, William R. “Selecting the Right Similarity‐Scoring Matrix.” Current Protocols in Bioinformatics 43 (1) (2013).
16. “PIR-PSD Database [PIR - Protein Information Resource].” https://proteininformationresource.org/pirwww/dbinfo/pir_psd.shtml (Accessed October 23, 2026).
17. Stevens, Hallam. 2013. A Data-Driven History of Bioinformatics. Chicago: University of Chicago Press, 2013.
18. Strasser, Bruno J. “Collecting, Comparing, and Computing Sequences: The Making of Margaret O. Dayhoff’s Atlas of Protein Sequence and Structure, 1954–1965.” Journal of the History of Biology 43 (4) (2009): 623–60.
19. Strasser, Bruno J. “The Experimenter’s Museum: GenBank, Natural History, and the Moral Economies of Biomedicine.” Isis 102 (2011): 60–96.
20. Zawilak, Tea, “Sanger Sequencing”. Embryo Project Encyclopedia (2025-08-08). ISSN: 1940-5030 https://hdl.handle.net/2286/R.2.N.202535