The Web of Science™ (Clarivate™, Philadelphia, PA, USA), the most trusted global database, and the strongest search engine, was used for publication retrieval on 25 April 2022. The retrieval strategy was as follows: (((TS=rheumatology) OR (TS=rheumatic disease) OR (TS=rheumatism)) AND ((TI=transcriptomic) OR (TI=proteome) OR (TI=proteomics) OR (TI=metabolomics) OR (TI=bioinformatics) OR (TI=metagenomics) OR (TI=metatranscriptomics) OR (TI=omics) OR (TI=microarray) OR (TI=sequence) OR (TI=RNA-seq) OR (TI=sequencing) OR (TI=ATAC-seq) OR (TI=single cell sequencing) OR (TI=single cell sequence) OR (TI=single cell RNA sequencing) OR (TI=single cell RNA sequence) OR (TI=expression profile) OR (TI=bioinformatic*) OR (TI=high throughput))). We retrieved a total of 1,374 articles published from 1 January 2000 to 25 April 2022 in the Web of Science Core Collection. Reviews and monographs were excluded as publication types, and articles, meeting abstracts, proceedings papers, editorial material, letters, and corrections were retained ( Figure S1 ). Altogether, 1,374 publications of these types were considered as relevant articles in this bibliometric study.

All data were exported in a text file and were analyzed using Bibliometrix (version 3.2.1), an open-source tool developed by Massimo Aria and Corrado Cuccurullo using statistical computing and R language (version 4.2.0, Institute for Statistics and Mathematics, Vienna, Austria; www.r-project.org) for quantitative research in scientometrics. Bibliometrix can support the workflow for bibliometric analysis and can be easily upgraded and integrated with other R packages (The R Foundation for Statistical Computing, Vienna, Austria) (27). Other than Bibliometrix, Citespace (5.8.R3) and VOSviewer (v.1.6.15, Centre for Science and Technology Studies, Leiden University, Leiden, the Netherlands) were also widely applied in this bibliometrics study. Developed by Dr Chaomei Chen using Java, CiteSpace can construct and assess co-citation networks via visualization analysis (28). VOSviewer, capable of generating a co-occurrence matrix from keywords (29), was also very popular among researchers. In the present study, countries, authors, affiliations, co-cited publications, and co-occurrence of keywords were analyzed. The dimensionality reduction technique was used to visualize the conceptual structure of keywords to identify the frontier studies and hotspots.

Since the year 2000, publications in rheumatism sequencing have shown an overall tendency of growth. From 2000 to 2004, few articles were published in this area. In 2005, the number of publications increased in line with the development of proteomics and maturation of microarrays. After 2009, with the emergence of single-cell sequencing, publication numbers rapidly increased, with more effort dedicated to the field using this technology. The data retrieval date (April 25) may account for the steep decline in 2022 ( Figure 1 ). Taken together, these results indicate that this field is attracting more and more attention and has great potential for future development.

A total of 1,374 articles were retrieved, from 62 countries, led by the USA and China, which accounted for 334 publications and 300 publications, respectively, creating a huge production gap between these two countries and the rest of the world. Ranked by accumulated publication frequency, the USA came first, with 1,592 publications, and China second, with 982 publications ( Figure 2A ). In addition, China and the USA were the two most cited countries across all publications, indicating excellence in both the quantity and quality of their research ( Figure S2A ). The top 20 most productive countries are listed in Table 1 , with the USA, China, Japan, Germany, South Korea, United Kingdom, Australia, Sweden, Italy, and India making up the top 10. Collaboration strength between countries can be inferred from the single-country publication (SCP) and multiple-country publication (MCP) rates. The countries that had the highest MCP ratio were Switzerland, the United Kingdom, the Netherlands, Canada, and Norway, that is, mostly European countries; other countries, such as China, Korea, and Belgium, focused more on domestic publication. As shown in Figure 2B , the USA had strong cooperative bonds with China, Korea, Canada, and European countries in particular, whereas China has strong bonds with the USA, Europe, and Korea. Although the USA’s and China’s MCP ratios were not the highest, they still ranked first and second, respectively, for total number of MCPs ( Figure S2B ).

During the last 22 years, a number of authors have played active roles and made great efforts to push the field forward. Lotka’s law describes the pattern between authors and their number of published papers (30). According to Lotka, the number of authors with n published articles is 1/n ² of those with one published article (31). In general, this law describes the phenomenon that a small number of authors produce a large proportion of relevant articles. This law also applies to publications in rheumatism and sequencing, as a large proportion of the publications in rheumatism and sequencing were published by a rather small proportion of authors ( Figure S2C ). The top 20 most productive authors are listed in Figure 2C , with Li, Liu, Baechler, Kumar and Misra accounting for the top 5. Judging by the h-index, Baechler and Klareskog were the most productive authors, with the highest H-index, of 8, meaning that they produced eight documents, all with an impact factor above 8. They were followed by Liu, Utz , and Dai, with an h-index of 7 ( Figure 2D ). However, the number of studies produced by Baechler reached its peak in 2005, and he ceased to produce studies altogether after 2017, whereas Kumar, although starting late, in 2015, has retained an energetic role until today. Li, Liu, Kumar, Misra, Blanco,Guleria, Takeuchi, Wang, Dai, Dubey, Liu, and Mariette continued to be active after 2020. ( Figure S2D ).

The source of authors’ inspiration, and the incubator for their discoveries, is their institution. Thus, it is of great importance to identify the prolific institutions. Figure 2E shows that Stanford University was positioned first, with 56 publications, followed by the University of California, San Francisco (n = 48), Brigham and Women’s Hospital (n = 41), the Karolinska Institute (n = 34), Leiden University (n = 34), and the University of California, San Diego (n = 34), four of which are in the USA. Shanxi Medical University, the affiliation of the most prolific author, Li, and the University of Minnesota, the affiliation of the third most prolific author, Baechler, also made the top 20. These results are in line with the accumulated production figures for countries noted above, indicating that the USA is in the vanguard of the field.

Since the year 2000, 1,374 articles have been published, from 350 sources. The top 20 most popular journals are illustrated in Figure 3A . Arthritis & Rheumatology (which was named as Arthritis & Rheumatism before 2014), Annals of the Rheumatic Diseases, and Arthritis & Rheumatism were the top three, all with publications numbers surpassing 140, accounting for the majority of the relevant articles ( Figure S3 ). During the last 22 years, cumulate co-occurrences in all journals presented an increase, among which Arthritis & Rheumatology and Annals of the Rheumatic Diseases displayed a rapid increase after 2012 or 2013, signaling an upsurge in the importance [or popularity of these journals. Arthritis & Rheumatism, on the other hand, manifested a stable level after growing from 2004 to 2013. Detailed trends are presented in Figure 3B . Furthermore, taking the impact of journals into account, Arthritis & Rheumatism ranks in first place, with an H-index of 19, followed by PLOS ONE (H = 18), Annals of the Rheumatic Diseases (H = 17), and Arthritis & Rheumatology (H = 16), making them valuable sources for research related to rheumatism and sequencing ( Figure 3C ).

Certain documents or references cited laid served as future directions for research into rheumatism. Globally cited documents are those that are cited in the Web of Science database, whereas locally cited documents are those that are cited in this collection of articles. Thus, high numbers of globally cited documents indicated overall influence, whereas high numbers of locally cited documents suggest an impact in this field. Figures 4A , B show the top 20 globally cited documents and the top 20 locally cited documents. The most globally cited document and fourth most locally cited document, “Complete genome sequence of an M1 strain of Streptococcus pyogenes” (32), by Ferretti in 2001, revolutionarily identified the genes responsible for “molecular mimicry” causing rheumatic fever or acute glomerulonephritis, and set an inspiring premise for subsequent research. It was cited in the paper “Genome sequence and comparative microarray analysis of serotype M18 group A Streptococcus strains associated with acute rheumatic fever outbreaks” (33) by Smoot in 2002 ( Figure 4C ), which is also the second most globally cited document, digging deeper into the molecular basis of acute rheumatic fever pathogenesis. “Peripheral blood gene expression profiling in rheumatoid arthritis” (34), by Batliwalla, published in 2005, was first among the most locally cited documents, and was also in the top 20 most globally cited documents, providing the groundwork of gene expression patterns of rheumatoid arthritis. “Gene expression profile analysis of rheumatoid synovial fibroblast cultures revealing the overexpression of genes responsible for tumor-like growth of rheumatoid synovium” (35), by Watanabe, published in 2002, “A 588-gene microarray analysis of the peripheral blood mononuclear cells of spondyloarthropathy patients” (36), by Gu, published in 2002, and “Discovery of distinctive gene expression profiles in rheumatoid synovium using cDNA microarray technology: evidence for the existence of multiple pathways of tissue destruction and repair” (37), by van der Pouw Kraan, published in 2003, were three pioneering articles in the exploration of rheumatoid pathogenesis. Among the most locally cited references ( Figure S4A ), “The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis” (38), by Arnett, ranked first, revealing the central and fundamental role of the definition and classification of rheumatism. The dominant position of this article can also be seen in the Sankey diagram ( Figure S4B ). Ranking the second in the most locally cited references was “Interferon-inducible gene expression signature in peripheral blood cells of patients with severe lupus” (39), by Baechler, published in 2003, which widened our knowledge of SLE. Baechler's work showed a high level of inheritance, that is, novel findings were derived from his own previous research.

Factorial analysis in bibliometric studies can identify a smaller group of factors that describe the correlations between the input literature, and form clusters according to the interrelationships among them (40). Articles that appear closer on the figure share stronger correlations. Applying factorial analysis, these most cited documents can be broken down into distinct clusters. In consonance with the historical direct citation network, the work of Smoot distributed in the vicinity of Ferretti, confirming their strong interrelationship ( Figure S5A ). As shown in in Figure S5B , the documents with the highest contributes, “Perturbations in amino acids and metabolic pathways in osteoarthritis patients determined by targeted metabolomics analysis” (41) by Chen, “Comparative analysis of synovial fluid and plasma proteomes in juvenile arthritis—Proteomic patterns of joint inflammation in early stage disease” (42) by Gibson, “Proteomics analysis for verification of rheumatoid arthritis biomarker candidates using multiple reaction monitoring” (43) by Lee, and “Proteomics in rheumatology” (44) by Kabeerdoss, scattered near each other. These four documents share a common focus, that of proteomics, demonstrating the significance of proteomics in the study of rheumatism. Relatively far from the clusters lies the work of Xie, “Bioinformatics analysis of epigenetic and SNP-related molecular markers in systemic lupus erythematosus” (45), which might indicate the emergence of epigenetics in rheumatism research.

Keywords that constantly appeared in selected articles reflect the hotspots for research to some extent. KeyWords Plus (Clarivate) are words or phrases that are calculated by special algorithms that do not appear in the title of the article but appear in the titles of the article’s references. Article keywords, in contrast, are supplied by the article authors themselves. The KeyWords Plus analysis ( Figure 5A ) revealed that “disease” and “expression” were the two most common keywords among the articles we examined with 112 and 110 occurrences, respectively, far exceeding all other keywords, implying that the field was disease oriented, and focused on dynamic gene expression instead of just sequencing. This high frequency of appearance of “disease” and “expression” was visualized using a keyword tree and word cloud ( Figures S6A, B ). Based on our analysis of authors’ keywords ( Figure 5B ), it is evident that most work was dedicated to rheumatoid arthritis, followed by rheumatism, osteoarthritis, and systemic lupus erythematosus.

In addition to determining the frequency of keywords, it is also of great importance to investigate their evolution and development. Over the period from 2000 to 2022, all selected keywords displayed an increasing level of occurrence. Among all keywords, “disease” and “expression” showed robust growth, indicating promising future development ( Figure 6A ). The Sankey diagram of keywords in which time is divided into the periods 2000–2014, 2015–2018, and 2019–2022 ( Figure 6B ) shows that “disease”, “identification”, and “rheumatoid-arthritis” attracted the most attention during the first decade of the century. The focus then shifted to keywords such as “expression”, “cells”, “cancer”, “gene”, “susceptibility”, and “pathogenesis”. However, in the most recent years, that is from 2019 to 2022, “disease” and “identification” regained their popularity among researchers. In addition, “rheumatoid-arthritis”, one of the best-known rheumatic diseases, has consistently been a focus for researchers over the last 22 years. It can be deduced from this diagram that sequencing technology mainly served as tools for the study of rheumatic diseases and their mechanisms. The prevalence of the word “expression” suggests the growing attention paid to transcriptomes, in turn reflecting the increasing value of sequencing technology. The emergence of “cells” may imply that is becoming associated more closely with studies investigating rheumatism on a cellular level is becoming associated more closely with studies investigating rheumatism on a cellular level. The value of exploring novel biomarkers via sequencing technology is confirmed by the emergence of “identification”.

In addition to identifying the popular words, it is crucial to analyze the co-words network to make out the conceptual structure. Keywords were divided into five clusters, as illustrated in Figure 7A , and words belonging to the same cluster were given the same color. The blue cluster mainly consisted of the terms “disease”, “association”, “susceptibility”, “risk”, “arthritis”, “systemic-lupus-erythematosus”, and “genome–wide association”. The red cluster included the terms “expression”, “identification”, “gene-expression”, “protein”, “cartilage”, “articular-cartilage”, “gene-expression”, and “gene”. The orange cluster was composed of the terms “cells”, “inflammation”, “activation”, “pathogenesis”, and “apoptosis”. The green cluster comprised the terms “classification” and “disease-activity”. The purple cluster contained the words “rheumatoid-arthritis”, “diagnosis”, “osteoarthritis”, “biomarkers”, “serum”, “mass-spectrometry”, and “synovial-fluid”. Diameters of the nodes were in proportion to the occurrences, and the line thickness that bridged nodes together indicated co-occurrences.

Applying factorial analysis, documents on rheumatism and sequencing could be divided into two major categories, indicated by the blue and red clusters ( Figure S7A, B ). The blue cluster, in which the central word, “biomarker”, is surrounded by the terms “mass-spectrometry”, “proteins”, “plasma”, “synovial fluid”, and “serum”, symbolizes a biomarker-centered topic, in which mass spectrometry is applied to look for novel proteins in plasma, synovial fluid, or serum. The other cluster, colored red, with the word “disease” in the center and surrounded by “criteria”, “classification”, “arthritis”, “association”, “mechanism”, and so on, refers to the diverse but relatively tightly linked aspects of rheumatic diseases.

Centrality refers to the relevance of certain themes to the field, and density refers to the degree of development of these themes. Five groups of themes were scattered on the four-quadrant graph with each quadrant representing different state of themes. Motor themes refer to the ones with high relevance and rapid development, serving as driving force in this field. Basic themes refer to the ones with high relevance and low development, serving as cornerstones in this field. Niche themes refer to the ones with high development and low relevance. Emerging or declining themes refer to the themes with low relevance and low development, which are probably the newly emerged themes or the themes losing its importance and attention ( Figure 7B ). The blue-colored group, comprising the themes “association”, “classification”, and “disease activity”, and the red-colored group, comprising the themes “expression”, “cells”, and “activation”, lay in the motor themes quadrant.. Themes such as “identification”, “rheumatoid arthritis”, and “protein” had solid connections to the filed but lacked development, serving as basic themes. Studies of themes related to “disease”, “risk”, and “susceptibility” were highly developed, despite their relatively weak association with the field of rheumatism and were regarded as niche themes. The themes of “diagnosis”, “biomarkers”, and “serum” and the themes of “senile idiopathic arthritis”, “peripheral blood”, and “I interferon” were placed in the quadrant of emerging or declining themes.

Although studies into rheumatism can be traced back thousands of years, our understanding of the disease was, until recently, severely hampered by its intricate characteristics and lack of efficient tools available for the exploration of its pathogenesis. Complicated clinical manifestations, complex categorizations, and a veiled pathogenesis, which could only be better analyzed with revolutionary technologies, stood in the way of our explorations. Former studies were restricted to analyzing rheumatism on the clinical and semiological levels, without a detailed analysis of the molecular level. Moreover, the critical role of pathogens in rheumatism, such as Streptococcus pyogenes in rheumatic fever, also demanded analysis from the molecular level. Fortunately, sequencing technology was developed and brought answers to these long-unsolved problems, as well as drastic changes to the field, as a result of its capability of investigating molecular patterns. Details of the pathological and immunological processes underlying rheumatic diseases were elucidated, more comprehensive disease classifications were designed, and hidden stages of diseases were revealed. Today, sequencing technology undoubtedly serves as a major driving force in studies of rheumatism.

Moreover, the rapid changes in sequencing technology are also altering the trajectory of rheumatism studies. The announcement, on 26 June 2000, that scientists from the Human Genome Project (HGP) had achieved the important milestone of completion of a working draft sequence, ushered in the post-genomics era, which placed more stress on the study of functional genomics, specifically the dynamic genome, and its products such as RNA, proteins, and metabolites (46). In line with this development of sequencing technology, which shifted from DNA sequencing to transcriptomics, then to proteomics, to metabolomics, and to epigenetics, sequencing technology enabled scholars to find genetic patterns of rheumatic susceptibility, its gene expression characteristics, and to apply proteomics to identify novel biomarkers.

In this bibliometric analysis, we examined 1,374 articles published from 2000 to 2022 (April 25) on sequencing technology and rheumatism.

The number of annual publications, to a certain extent, can reflect the level of researchers’ interest in the field. The number of articles on rheumatism and sequencing started to increase significantly after 2000 and has continued to exhibit an overall increase, despite a few ups and downs. In the year 2005, the number of publications saw its first peak, with most publications focusing on proteomics and gene expression profiling. The same theme lasted for years, until the technique of single-cell RNA sequencingbecame mature and was widely deployed, as a result of which it became the most popular research topic in rheumatism and sequencing. Sequencing in rheumatism regained its popularity in 2009, and the number of articles published increased to nearly 100 per year. In 2017, the publication number received another boost, and in 2021 reaching a record high of 152 publications.

Among the 62 countries that dedicated efforts to this field, the USA ranked first, in terms of both the total number of documents and total number of MCPs, and was the home to a large proportion of most relevant affiliations, such as Stanford University, the University of California, San Francisco, and so on. China came in second in total publication numbers, but had the highest number of SCPs. Countries also worked together to conquer obstacles, with the USA having the strongest cooperative bonds with China, European countries, and Korea. Switzerland had the highest MCP ratio, indicating that most of its publications resulted from collaboration with other countries. In terms of the most relevant authors, Li from Shanxi Medical University and Liu from the First Affiliated Hospital of Anhui University of Traditional Chinese Medicine were at the top, with 16 and 15 related articles published in total, respectively, and they are still very active today. Baechler of the University of Minnesota and Gregerson of the Feinstein Institutes for Medical Research reached their peak publishing activity in 2005, and their output has since gradually declined. Several connections can be found between the most relevant authors and affiliations. For instance, Shanxi Medical University, which ranked 12th on the most relevant affiliations, is the institution where Li, the most prolific author, works. This indicates that the most relevant authors also played prominent roles in the establishment of great teams and research environments in this field.

Arthritis & Rheumatology, Annals of the Rheumatic Diseases, and Arthritis and Rheumatism were the three most relevant sources, among which the first two presented an increase in 2013. Although PLOS ONE had only an impact factor of 3.752 in 2022, it published 36 relevant articles and had an h-index of 19, so its influence on the subject should not be ignored.

Most cited documents, whether globally cited or locally cited, served as milestones for crucial discoveries and as important stimuli of research. For instance, “Complete genome sequence of an M1 strain of Streptococcus pyogenes” (32), by Ferretti, published in 2001, mapped out the complete genome sequence of a strain of the pathogen that is responsible for rheumatic fever. This work has greatly increased our understanding of rheumatic fever, which still remains uncontrolled in some countries and regions (47). This work also motivated the study “Genome sequence and comparative microarray analysis of serotype M18 group A Streptococcus strains associated with acute rheumatic fever outbreaks” (33), by Smoot, published in 2002. Another perspective on the human immune system response to rheumatic disease is provided by the article “Interferon-inducible gene expression signature in peripheral blood cells of patients with severe lupus” (39), by Baechler, published in 2003. This article, in which Baechler identified IFN pathway dysregulation as a marker of more severe disease in SLE patients, played a central role in the field of SLE research, not only cracking the pathogenesis of SLE but also providing insight into the treatment for it. Although this article was not on the list of either the most globally or locally cited documents, it ranked second on the most locally cited references, revealing its significant impact. Together with “Microarray analysis of microRNA expression in peripheral blood cells of systemic lupus erythematosus patients” (48), by Dai, published in 2007, the third most globally cited document, it is apparent that the peripheral blood acts as an essential substrate for studying the pathogenesis of SLE. From this analysis, two lines of inquiry could be identified in these publications. One focused on the pathogens that might lead to rheumatic disorders, while the other focused on exploring the inner characteristics and immunological responses of rheumatic diseases. Altogether, it can be seen that the focus in the field was transferred from genome sequencing to gene expression profiling and to proteomics. This trend synchronizes with the central dogma of molecular biology and the advancement of sequencing technology.

“Disease” and “expression” were the most frequent keywords identified. In addition to these, “rheumatoid arthritis”, one of the most popular diseases studied in rheumatism, accounted for 5% of total word occurrences. Keywords concerning pathogenesis, such as “inflammation”, “activation”, “cells”, and “pathogenesis” itself, were also high in frequency. “TNF-alpha”, "NF-kB", and “tumor necrosis factor” could also be seen in the word cloud, suggesting that there was an emphasis on the study of immunological molecules. In the co-word analysis, keywords were divided into five distinct clusters.