Anthropogenic activities disseminate large amounts of chemical substances into the environment [1]. Hence, humans are currently exposed to several pollutants with genotoxic potentials such as metals [2], pesticides [3], industrial waste mix [4], and nanomaterials [5]. Genotoxicity is a wide term comprising DNA damage and mutagenicity, where the mutagenic effect is described as an occurred event with irreversible and heritable outcomes affecting the DNA and/or chromosome structure [6]. The genotoxic effect recorded on somatic cells has been associated with pathological endpoints such as premature aging, neuronal diseases, and even carcinogenesis [7,8]. Accordingly, the genotoxicity evaluation is a required component in the human health risk assessment and as one single test is unable to detect all the genotoxic endpoints, a battery of in vitro and in vivo tests was recommended [9]. Among these assays, the somatic mutation and recombination tests (SMARTs) are one of the commonly used tests. The SMART assays are in vivo assay to assess the potential genotoxicity of substances in the somatic cells of Drosophila melanogaster [10]. This assay could target wing cells named as wing-spot test proposed first time by Graf et al. [11], or eyes cells known as eye-spot test defined by Wurgler and Vogel [12]. In both cases, losing heterozygosity by deletions, point mutations, mitotic recombination, and nondisjunction unravels the expression of genetic markers in heterozygous or transheterozygous individuals, ensuring the quantification of the damage by visual scoring [13]. The wing-spot test includes two types of cross using recessive genetic markers on the 3rd chromosome, a standard cross with normal bioactivation between female virgins (flr³/In (3LR)TM3, ri p^psep l(3)89Aa bx^34e e Bd^s) and (mwh/mwh) males, and a high metabolic bioactivation cross with high levels of cytochrome P450 between female virgins (ORR; flr³/In (3LR)TM3, ri p^psep l(3)89Aa bx^34e e Bd^s) and (mwh/mwh) males. The mutant spots are produced after chemical exposure that induced point mutation, deletion, or mitotic recombination [14,15]. The eye-spot test is based on a cross between wild-type eyed females (w+/w+) and white-eyed males (w/Y). The gene white (w) is a recessive marker found on the X chromosome. During the offspring period, a mutagenic event could occur and cause the formation of white phenotype spots (mutant ommatidia) in the wild-type eyes [16]. Although both tests are accurate, sensitive, and specific, the wing-spot test allows the visual scoring of wings over time, whereas in the eye-spot test, the analysis should be performed quickly since no preserving actions are available on the eyes [17]. In fact, the SMART assays have been applied in the analysis of the genotoxicity and the antimutagenicity of several chemicals and agents such as pesticides [18-20], nanomaterials [21,22], food products [23,24], hormones [25], plant extracts [26], and drugs [27-29]. With this variety of studies, a bibliometric analysis is required to assess the impact of this methodology on the genotoxicity studies. Bibliometric analysis is considered a highly sensitive method to evaluate research outputs based on statistical tools and to study metrological features of information created in a specific field [30]. To the best of our knowledge, no paper using the bibliometric analysis to explore the trends of SMART assays research has been published. Therefore, in the current study, various aspects were exanimated to evaluate the publications and citation trends in SMART assays from 1984 to 2020. Hence, the following research objectives were considered guiding the study design: (1) to identify the most influential journals and publications, the impactful authors and institutions, and the leading countries in SMARTs literature; (2) to find the patterns of collaboration between countries and authors within this research domain; (3) to explore the emerging keywords and research themes. The findings of the present study will provide a comprehensive overview of the importance of SMART assays, it would provide information to the scholars to easily identify the research profile and enhance collaboration.

This bibliometric study analyzed the published academic studies of SMARTs indexed in the Web of Science (WOS) core collection database. As known, WOS is the most reliable global citation database with a collection of over 21,000 peer-reviewed journals and the most accepted one for analysis of academic papers [31]. A comprehensive four-step approach was framed in this study as shown in Fig. 1. Boolean operators were used with adding all the relevant keywords to retrieve more relevant papers. (TS = ((“SMART assays” OR “SMART assay” OR “SMART test” OR “SMART tests” OR “wing-spot test” OR “wing-spot assay” OR “somatic mutation and recombination test” OR “eye-spot test” OR “eye-spot assay” OR “small single spots” OR “large single spots” OR “mutant ommatidia” OR “mutant eye unit” OR “w/w+ SMART”))). The search was performed on 14 January 2021; it is essential to present the date of records collection as the database is constantly updating [32]. Most bibliometric studies are based on academic articles [30], thereby this study was limited to original articles written in English in the strict sense. The search was also limited to 2020. The authors adopted the PRISMA approach, which has been used in bibliometric studies. A total of 460 records were extracted initially, which later were filtered by document types, to exclude the following document types: proceedings papers, book chapters, and early access. The data relevance and accuracy were assured by scanning the title and abstract of each record, and 49 records were excluded. The irrelevant articles were from the following categories: (1) Internet of things and wireless sensing, (2) communication services, (3) image processing, (4) industry, and (5) single molecule amplification and re-sequencing technology (SMART). Finally, a total of 392 records were remained to be downloaded in plain text to extract the following data: publication year, author, title, abstract, keywords, cited references, journal title, and institution, for further analysis.

Having selected 392 relevant records, data visualization and analysis were conducted by using various bibliometric software and applications, including Microsoft Excel, HistCite, Biblioshiny (RStudio), VOSViewer, and CiteSpace. The analysis was performed in three stages. First, a bibliometric citation analysis was performed (number of publications and citations, relevant journals, productive institutions, and the most impactful articles and authors). Secondly, a network analysis was applied, including collaboration between countries and co-authoring using the Walktrap clustering algorithm. This algorithm has the advantage to be computed effectively and placing the data in a network [33]. Furthermore, a three-fields-plot based on the Sankey diagram to indicate the interrelation between keywords, journals, and countries was visualized. Detection of the emerging research fields is required to outline research areas. Therefore, thirdly, a content analysis was conducted. HistCite (version 12.03.17) was used for the bibliometric citation analysis to sort the collected data by quantitative (number of publications and citations) and qualitative indicators (total global citation score and total local global citation score). Hence, these two types of indicators were applied in the current study. The total global citation score (TGCS) represents the number of citations of a paper included in the collection selected for the analysis in the WOS whereas the total local citation score (TLCS) refers to the total number of citations of a paper included in the collection and has been cited by other papers of the same collection [34]. The RStudio software (version 1.3.1093) with Biblioshiny application was used [35]. Hence, this application was also used to analyze the basic indicators of the search, plus the collaboration between countries and authors, and the three-fields plot. VOSViewer software (version 1.6.16) was used to analyze and visualize the emerging keywords. CiteSpace (version 5.7.R4) was used to perform the co-citation analysis and to investigate the emerging topics. Co-citation analysis is an effective tool to understand the intellectual structure of a research field, with the cited papers intellectual bases are revealed whereas papers in their active state of citation represent the research frontiers which usually display characteristics of a field specificity [36]. One of the important tools found in CiteSpace is the betweenness centrality. A paper with high betweenness centrality was defined as an important research paper in the network [37]. Furthermore, articles with strong citation bursts with time slices were identified. The burst is observed when a publication has an excess in its citation counts compared to its peers. This furthers identify publications that interested the scholars over time and thereby help to explore the research frontiers of a given field [38,39].

The present study analyzed the SMART assays in the genotoxicity studies published during 1984-2020. A total of 392 records have been written by 933 authors from 35 different countries with an average of 17.01 citations per document. Authors of single-authored documents are seven (0.75%) while authors of multi-authored documents represent 926 authors (99.25%).

In the current study, a comprehensive analysis of the emerging trends in the field of SMART assays from 1984 to 2020 was performed. The analysis reveals an increase in the number of publications in this period, with most of these having been published in the last 10 years, confirming the growing interest in SMART assays. Of note, recently published studies have received fewer citations compared to ancient studies, as it requires time for a study to make an impact. The first observed bloom of TLCS and TGCS has been associated with the first paper introducing the SMART assays in 1984. Thus, the highly cited article is the earliest publication. The second bloom in 1992 has been generated from two influential articles written by Graf and van Schaik (1992) [41] and Frei et al. (1992) [42]. The most prolific author in the SMARTs field is Graf U (Switzerland), with 953 TLCS and 1,716 TGCS. Furthermore, four of his research papers are in the top 10 highly cited articles, and five of them in the top 25 strongest citation bursts showing the interest of researchers in his field. The collaboration between authors could be due to the emergence of interest among researchers. While only one of the prominent clusters is interconnected, it is expected in the future to improve overall collaborative work. The large number of multi-authored documents could be related to various collaborations between countries to expedite the usage of SMART assays. In fact, 35 countries have published about SMART assays. The two influential authors de Andrade HHR and Lehmann M are both from Brazil. Thus, this country had the highest number of publications. The rise in publications from Brazil can also be attributed to the high frequency of collaboration with institutions in Switzerland. Both authors de Andrade HHR and Lehman M were collaborators with Graf U in 2000 to study the genotoxic potential of tannic acid [56]. However, the usage of SMART assays in genotoxicity studies is still ignored in many countries. The most influential journals accounted for 61.5% of all the publications, and this finding illustrated that the distribution of publication was narrow. These influential journals are in Q1(3), Q2(2), Q3(3), and Q4(2) category. To note, five highly cited articles were published in Mutation Research (Q1), two in Mutagenesis (Q2), and one in Environmental and Molecular Mutagenesis (Q2). The first most-cited article by Graf et al. was published in 1984 [11]. This article was published in Environmental Mutagenesis (Q2), ranked at 55th position with only this publication. Therein, the protocol of SMART assay was presented, and several chemicals such as β-propiolactone, 1,2-dibromoethane, aflatoxin B1, diethylnitrosamine, dimethylnitrosamine, mitomycin C, and procarbazine have been identified as mutagens. The second most-cited article by Frei and Wurgler [40] was published in Mutation Research-Environmental Mutagenesis and Related Subjects. This journal ceased publication, to be incorporated in 1997 into Mutation Research-Genetic Toxicology and Environmental Mutagenesis journal (Q3) [57]. In this study, to reduce the risk of inconclusive results, a new statistical test was proposed. The third most-cited article by Graf and van Schaik [41] proposed new strains to improve the visual score. This article was published in Mutation Research (Q1 in 1999) currently known as Mutation Research- Fundamental and Molecular Mechanisms of Mutagenesis (Q3) [57]. The most used keywords allow distinguishing the articles relevant to SMART assays. The three-fields plot showed that Drosophila melanogaster, genotoxicity, SMART, and antigenotoxicity were used by authors from Brazil, Turkey, and Spain. These keywords appear to be generic; yet it was used frequently. Based on thematic analysis, five large clusters were observed. The two clusters ‘somatic mutation’ and ‘nitrogen mustard’ are the two largest and oldest clusters. The cluster ‘copper oxide nanoparticle’ is the youngest cluster. The value of the mean silhouette (S) and the modularity (Q) are 0.9501 and 0.3944, respectively, suggesting reliable and robust results. In the largest cluster labeled ‘somatic mutation’ (#0) with 103 members and which is also the most active cluster, the most actively citing article (a research frontier article) identified polycyclic aromatic hydrocarbons and their derivatives as genotoxic [58]. The most actively cited articles (intellectual-based papers) are Graf et al. (1989) [14], Lindsley and Zimm (1992) [48], and Graf and van Schaik (1992) [41]. Graf et al. [14] is the second top-ranked paper with the strongest citation burst (11.22) and therein the efficiency of the SMART assay was proven by applying it on 30 compounds to evaluate their genotoxic potential. Lindsley and Zimm [48] described and identified the genome of Drosophila. Graf and van Schaik [41] improved new strain to assure better cross. In brief, cluster #0 mainly concentrated on enhancing the SMART protocol to be more practical and accurate. In the second cluster labeled ‘nitrogen mustard’ (#1), the most actively citing paper focused on the effect of tannic acid on nitrogen mustard, mitomycin C, and methylmehanesulfonate [56], whereas the most actively cited paper are Frei and Wurgler [40] and Graf [59]. These two papers focused on the Drosophila model by identifying the sample size required as well as the appropriate age of larvae, depending on the mutagens to avoid inconclusive results. It can be concluded that cluster #1 focused on the incorporation of a new control positive (nitrogen mustard) into SMART assay. Not surprisingly that the two largest clusters cover the most interests as both focused on the improvement of SMART assays protocol. In the third cluster labeled ‘vivo model’ (#2) the most actively citing paper was written by Carmona et al. [60] to evaluate the genotoxic potential of titanium dioxide anatase nanoparticles. Demir et al. [49] and Vales et al. [52] are the most-cited paper. Demir et al. [49] analyzed the genotoxic potential of silver nanoparticles whereas Vales et al. [52] evaluated the genotoxic effect of cobalt nanoparticles. It seems that in cluster #2 nanoparticles have gradually attracted the attention of scholars. Although the most-cited and citing papers were mainly dedicated to the genotoxic effect of nanoparticles, most of members treated various compounds, justifying thereby the label ‘vivo model’ instead of nanoparticles. In the fourth cluster labeled ‘doxorubicin-induced somatic mutation’ (#3), all the intellectual bases are related to cancer. Fragiorge et al. [53] analyzed the antigenotoxicity effect of ascorbic acid on doxorubicin. Doxorubicin is an antibiotic to treat human cancers which generally induces genotoxicity by oxidative damage [61]. Bishop and Schiestl [54] studied the function of homologous recombination in cancer. Similarly, the citing paper focused on the use of the herbal extract of ginseng to inhibit the genotoxic effect of doxorubicin [62]. In brief, cluster #3 was concentrated on resolving the doxorubicin-induced genotoxicity. The fifth cluster labeled ‘copper oxide nanoparticle’ (#4) is considered a persistent cluster denoting the continuity of an existing trend. In this cluster, the research frontier assessed the effect of copper oxide nanoparticles [63]. The metal oxide nanoparticles possess a redox property suggesting an antigenotoxic and anticarcinogenic potential [21]. The intellectual base emphasis on nanomaterials. Alaraby et al. [64] discussed the side effect of nanomaterials. Carmona et al. [60] presented the titanium dioxide anatase nanoparticles, being generally used in pharmaceuticals and cosmetics. This nanomaterial can generate oxidative stress, and subsequently genotoxicity [65]. Interestingly, this paper was a research frontier in cluster #2. It is worth mentioning, the first top-ranked burst item Frei and Wurgler (16.12) [47] is in a small and old cluster (1986) labeled ‘antiparasitic nitrofuran’ (#6) with 45 members. In sum, the research frontiers share sometimes the same theme with the intellectual bases since most often the continuation and the growth of intellectual base are the research fronts. The present study has certain limitations. Data were extracted from the WOS database; other databases were not considered. The database WOS is always updating, even with recently published papers, top publications are high enough that including recent paper would not have an influence on this research. The proceedings papers, book chapters, and early access are not included. Even though book chapters were excluded, the book by Wurgler and Vogel (1986) [12] describing the eye-spot assay was not indexed in the WOS database. The researchers tried carefully to include a maximum of relevant keywords; however, two studies Martinez-Valdivieso et al. (2017) [66] and Fernandez-Bedmar and Alonso-Moraga (2016) [67] have been missed. The absence of these papers is due to keyword search where none of the relevant used keywords were present in the title, neither in the abstract and/or the keywords section. The number of authors may differ since some authors published articles with a different initial of the first name. This led to some authors being separated into two authors. Even with these limitations, the current study on SMART assays can provide a directive to researchers to find influential articles, journals, and authors to assure collaboration and to reduce the research gaps.

In this study, a comprehensive bibliometric analysis was performed on SMART assays literature. Most of the publications were published in the last 10 years, mainly from Brazil, especially from the Federal University of Uberlândia. Notably, multiple authors were publishing papers on this field and mostly in ‘Food and Chemical Toxicology’ and ‘Mutation Research Genetic Toxicology and Environmental Mutagenesis’ journals. By analyzing the emerging trends, the nanoparticle area represents a persistent cluster. This area is expected to draw more attention. Finally, furthers studies should assess the records present on other databases to find out if the same trends for SMART assays are present.