The evolution of knowledge in sericultural research as observed through a science mapping approach

Sericulture, since its discovery in China, has spread to Background become a valued activity in a range of other countries. China remains the leading producer of silk, followed by India and other Asian countries, Europe, Brazil and Colombia. This article examines the evolution of sericultural research between 1892 and 2016, identifying the main themes and applications. : The SciMat software tool and the Bibliometrix R package were Methods used as tools for data analysis in this study, based on records from the Clarivate Analytics Web of Science and SCOPUS databases. : The results show that research has been growing, both in number Results of publications from the 1990s onwards, and in the emergence of topics closely related to the sericulture research field, a field that proves to be multi-disciplinary, exhibiting expansion and vitality. The information gathered will contribute greatly to the Conclusions: definition of relevant research strands, bearing in mind a number of significant gaps in information in this field. It will furthermore provide a better insight into the development of sericulture research over time.


Introduction
Sericulture is an agro-industrial activity that involves mulberry (Morus spp) cultivation, silkworm rearing, and the collection of silk thread for the textile industry. The latter aspect has been highly exploited 1,2 . Sericulture is carried out mainly in rural and suburban zones. Since it is limited by the supply of mulberry leaves, a lot of plants are required near the breeding site. It is environmentally friendly because it does not generate any pollutant waste, smell or sound and can therefore coexist happily with already populated areas. Silkworm rearing is not recommended however for industrial or intensely agricultural zones that use agrochemicals, given that silkworms are seriously affected by the presence of toxic product 3,4 .
Silkworms feed on fresh mulberry leaves (leaves and young sprouts) in which they find the nutrients and water necessary to grow and develop. While artificial diets do exist, they are used only in native breeding silkworm countries, such as Japan, because of their high cost 3,5,6 .
Mulberry was first cultivated for sericulture in Asiatic countries approximately 4,500 years ago. The practice of sericulture prompted the movement of species and varieties of mulberry plant across the continents, distributing these woody plants everywhere from temperate to tropical and subtropical climates 2,7 . Consequently, they do not have a precise origin. Many authors agree however that the main sites of origin include regions of China, Japan and the Himalayas 8,9 . Mulberries can grow on nonfertile ground, but when cultivated on rich ground, with regular irrigation, they produce a large quantity of high quality leaves.
In the textile industry, no natural product surpasses the quality of silk thread, because it possesses unique characteristics in resistance, elasticity, durability, and refinement. Silk thread can also preserve heat, absorb water, gases, and colorants 3,10,11 ; these are characteristics appreciated by the international market due to a rising tendency to use natural products, meaning that silk is increasingly welcomed and in demand.
Sericulture is an enterprise in which family members can work at a handcraft level so it requires low initial investment and less staff. Likewise, sub-products derived from thread production such as silkworm chrysalis and mulberry leaf residues can be used as feed for livestock, generating additional income for families.
China is considered to be the first country where spinning and knitting of silk thread took place, according to 7,000 year old archaeological findings in Yuyao city 12 . The trading of silk thread permitted an extensive, permanent and active economic interchange between China and other countries in ancient times, with the "Silk Road" connecting China, Mongolia, India, Persia, Arabia, Syria, Turkey, Europe and Africa 12 . Nowadays, spinning and knitting of silk thread is carried out in 30 countries, generally by small farmers. The main producers are China and India, who possess more than 50% of worldwide production, followed by Japan, Korea, and Thailand. The main producers in Europe are Italy and Spain, Zambia and Uganda in Africa, while in Latin America, Brazil is the largest producer, with Bolivia and Colombia as smaller producers 13,14 . Main consumer countries are the United States, Italy, Japan, India, France, China, United Kingdom, Switzerland, Germany, the United Arab Emirates, South Korea, and Vietnam, mainly for use in the production of clothing and home decor pieces 15 .
Besides in the production of silk for use in textiles, it is foreseen that sericultural research will be involved in using mulberry as a bioactive substance, able to function as medicine or food, where the silkworm is studied as a "bio-factory insect" for producing protein, thanks to transformation of the mulberry. The silk itself, due to its proteic nature, could be used to manufacture membranes, as a foodstuff, in cosmetics, or for medical and bioengineering purposes.
We sought to carry out a bibliometric analysis of research into sericulture in order to identify the principal topics related to sericulture research and their applications.

Methods
General methodology and description of the software tools used We followed four main steps to carry out the bibliometric analysis: 1. Datasets from different sources were extracted; 2. The data was cleaned up; 3. A particular analysis of the data was carried out, depending on the software tool used; 4. Lastly, the results were analyzed and discussed.
SciMAT 20 and the Bibliometrix 21 R package (version 3.3.3) were used as tools for data analysis in this study. Both are used widely for bibliometric analysis studies. SciMAT is widely used for management and organization 22,23 , virtual and remote labs 24 , knowledge-based systems 25 , marketing, and 26,27 in the field of social work. Bibliometrix is commonly used in businesses, healthcare, and public administration 28,29 .
SciMAT is based on the science mapping analysis approach presented in 20, and allows us to carry out science mapping studies within a longitudinal framework 30,31 . The SciMAT approach establishes four steps for analysis 20 : 1. Detection of substructures, especially clusters of words, by means of co-word analysis over specific time periods 2. Generation of cluster diagrams based on results of the first step.
3. Generation and interpretation of evolution diagrams based on clusters and periods. This reveals the general evolution of the research field. 4. Analysis of the different time periods, clusters, and subject areas that have evolved using the various measures of density and centrality generated by the tool.
A strategic diagram represents the themes of the research field. Classification into four groups is achieved using the method proposed in 32 and 33 for equivalence index, normalized for a co-word network of keyword co-occurrence. The Simple Centres algorithm is based on 34 and the values of cluster centrality and density rank are based on 32. Based on centrality and density, the clusters can be classified into four groups: An analysis of the evolution of research themes was also undertaken to detect the general areas of evolution, based on a map.
Bibliometrix is an R package (version 1.8) that comes with options to import bibliographic data from SCOPUS databases. The user can perform a science mapping analysis and build data matrices for co-citation, coupling, scientific collaboration analysis, and co-word analysis.

Datasets and search methodology
This study made use of two of the most important databases for scientific literature, i.e. the core collections of Scopus and Web of Science (Web of Science).
The analysis was performed using SciMAT version 1.1.03, configured with the following parameters: Unit of analysis, Words (authorRole=true, sourceRole=true, addedRole=true); Kind of network, Co-occurrence; Normalization measure, Association strength; Cluster algorithm, Simple Centres (Max cluster size: 7, Min cluster size: 3); Evolution measure, Association strength; Overlapping measure, Association strength. Some 1,883 records were obtained by Scopus and a further 312 by Web of Science, related to journal articles, notes, books, book chapters, reviews, and conferences. Following the deletion of duplicated records, the total of 2,195 was reduced to a final 1,930 records to be analyzed. On preliminary analysis, 91% of these were found to be for journals, with others totalling just 9%. We adopted a generalized definition of "Sericulture" which includes several different types of work on the silkworm (B. mori), mulberry (Morus spp), silk, and its industry. Some meaningless keywords that were too general, such as study or paper, were removed.
Using Bibliometrix as an R package does not require any special configuration. However, some steps based on R commands are needed: 1. Upload to R the download file in a character vector.
2. Convert the object to a data frame.
3. Apply the "biblioAnalysis" function that returns an object of class "bibliometrix". 4. Apply the "biblioNetwork" function with the "collaboration" parameter for the "analysis" and the "countries" parameter for the "network" variable. This creates a collaboration network that is then plotted.  period, but 93% of records were from 1990 onwards. According to these results, it could be inferred that sericulture is an increasing area of research. Of the 1,930 records found, 531 were from the years 1892 to 2000, and 1,399 from the years 2001 to 2016.

Data extraction
The most representative and influential journals that published on sericulture were mostly concentrated in Japan and India, except for Biomacromolecules, which is owned by the American Chemical Society.
The Journal of Sericultural Science of Japan and Japanese Journal of Human Geography are also listed, but to date these journals are not found in internationally indexed databases (Table 1).
Ten authors published 21.7% of the total research articles found, touching on various topics such as mulberry cropping, silkworm rearing, silk thread collection for the textile industry, and development of sericulture in different countries ( Table 2). These ten authors have an affiliation from institutions located in India. The country with the largest research output was India, with 1,242 records, much higher than the next two, Japan with 185 and China with 182. These three countries published 81.9% of the sericulture records found, and the ten most active countries produced 90% of the research (Figure 2).
Thirty-one countries gave meaningful contributions to sericulture from all over the globe. By continent, Asia is represented by 14 countries, Europe by 9, America by 5, Africa by 2, and Oceania by one. There are isolated publications from 19 countries, including Brazil, Italy, Turkey, Iran, Spain, and Colombia ( Figure 3). Other nations with a minimal number of records found in Web of Science include Croatia, Indonesia, Madagascar, Nepal, Pakistan, Paraguay, Romania and Russia.
Once the download references were uploaded in SciMat, a total of 2,518 keywords were extracted. Table 3 presents the top ten keywords identified. These keywords represent topics related to the concepts presented in Figure 4. This figure is based on the conceptual relationship between the keywords extracted, presented in clusters.     The number of keywords added in each of the four later time periods (64.9%, 60.5%, 48.8% and 58.5%, respectively) shows the vitality and expansion of the research field. Meanwhile, unchanged keywords between 38.9 and 53.1% (53.1%, 52.5%, 38.9% and 41.5%, respectively) point to the existence of certain base research topics ( Figure 6).
The conceptual evolution of sericultural research (Figure 7) is represented by green, red and blue colours, denoting mulberry cultivation, worm breeding, and silk, respectively. There are no gaps in the evolution of most of the topics. Seven topics (Protein, Larva BM, Silk, Silkworm, Genetic, India) have been present in the field across the five time periods analysed and may thus be considered classic themes.
The topics (mulberry cultivation, worm breeding, and silk) exhibit a positive pattern of development because of the increasing number of subtopics within them. For instance, in the case of Fiber, it diversifies into Sericin, Genetic, and Stress in the second period (Per2). Similarly, Cocoon (Per2) comes from BM larvae (Per1) and diversifies into Preservation, India, Parasite, and Economic-Conditions. Some topics sustain the number of topics from previous periods    and diversify into the next; this is the case for Microscopy (Per1) that comes from the initial period (Per0) of Cocoon and Protein and changes to Fiber and Stress in the following period (Per2) (Figure 7).

Research themes
To analyze the most prominent themes in the field of sericulture research for each time period, a strategic diagram is provided (Figure 8). In each diagram, sphere size is proportional to the number of records associated with each research theme.
Evolution of keywords related to sericulture over the last five years (2012-2016) The evolution of keywords in the last 5 years is depicted in Figure 9 and Figure 10. The relationships between research themes are presented in Figure 11. new keywords were mentioned. The addition of more than half the number of keywords in each period (60.5%, 62.4%, 57.2%, and 59.6%, respectively) and the fact that less than 45% of keywords remained the same (44.7%, 39.2%, 37.2%, and 28.8%, respectively) shows the vitality and expansion of this research field in the last five years ( Figure 10).
The conceptual evolution of the field of sericulture research in the last five years (Figure 12), in which the theoretical evolution      Figure S6).

Research themes in the last five years
To analyze the most prominent themes in the field of sericulture in the last five years, a strategic diagram is provided ( Figure 13).

Distribution of records in Scopus and Web of Science
Three general time periods were identified for publications: (1) Pre-1990s. Only 104 publications found, possibly because over that period publishing in annexed databases wasn't common, and the priority was other kinds of documents such as technical reports from the centres dedicated to sericulture research and journals with no annexed databases; (2) 1990s. Publications on sericulture begin to increase and specialized journals are created; The publication dynamics in sericulture appear to be related to the history of fibres. Ever since the appearance of synthetic fibres (early XX century) 16 , the market has evolved thanks to the cost, ease of production, greater uniformity, 17 , such that they tended to push natural fibers out of the marker. However, in the last few years, natural fibres have been viewed as a sustainable option, 100% biodegradable, environmentally friendly, this has caused their use to diversify and increase 18 .

Core sericulture journals
The journals that have published the most on the topic of sericulture have a low impact factor that can be attributed to their specialization. The appearance and indexing of specialized journals such as Silk and India Journal of Sericulture in the 1990s lead to an increase in publication output. The total number of journals that have published sericulture related topics is 528. The top 10 journals account for 62.4% of the total contributions (1,206).

Most productive authors
Publications in the last 20 years have been led by authors from India. Notably, the Central Sericultural Research and Training Institute, founded in 1961 and directed by the Ministry of Textiles of the Government of India, is actively developing the silk industry in the country, and its researchers represent the most productive authors. From the total number of authors (3,058), ten have published 21.7% of the research output (420 documents) and 83.5% (2,555) have published one or more documents.

Most productive countries
The importance of India in the production of sericulture documents worldwide may lie in the investment the country makes in developing the silk industry, including silk centres, research institutes, specialized journals, and organized events. Publications with collaborations are evident; Japan has the most publications in collaboration with authors from other nations with 12 country links, followed by China with 10 and India with 6 links. This might suggest that collaborative work is missing between countries with an established research output and those looking at sericulture as an option for diversification and economic development, as is the case with Colombia and Brazil in South America 19 .
Evolution of topics related to sericulture Generally, the keywords found were not linked just to one area of sericulture (mulberry cultivation, worm breeding, silk), showing that sericulture is multidisciplinary. It is worth mentioning that this research has shown that sericulture is dynamic and has faced new challenges created by environmental needs and scientific progress, as shown by keywords such as molecular, transgenic, genomic, sequence, and phylogenetic.

Conclusions
A bibliometric analysis of publications between 1892 and 2016, was carried out for the field of sericulture research. It was found that most research is published in journals in China and Japan, while the most representative authors were affiliated to research centres in India. As a result, India produces 64% of publications. The research field is centred on topics related to silk, silkworm, Bombyx mori, mulberry, cocoon, India, BM breeds, economic conditions, genetic, and development. As proposed by 35, B. mori is excellent as an experimental animal for genetic and biological research, and was present in several clusters in the course of the analysis. The field is multi-disciplinary and exhibits expansion and vitality. In future work, this type of bibliometric analysis gives researchers a broader view of the research field, to better identify potential interplay of topics in future work. Supplementary material Figure S1. Present topics inside the clusters of the last five years.

Data availability
Click here to access the data. Click here to access the data.   Click here to access the data. Despite the interesting topic, the reviewer has identified several weaknesses, including the need for description of data processing and more clear interpretation of results. Exact pre-processing steps and WOS/SCOPUS query extraction are essential to ensure the reproducibility. In addition, the review of literature on bibliometric analysis is highly recommended.
Several figures are not clearly described. For example, Figure 3 states "Collaboration between countries", whereas in the text it refers to isolated publications by country. It is also desirable to re-render figures with a better quality, if feasible.
More specific questions: Tools SciMAT tool -Cluster classification -consider adding explanation to four cluster groups In what follows, some comments and suggestion are listed. The introduction describes the sericulture research field deeply, but the objective and research questions of the paper are missing. Authors must to clarify if there are other bibliometric analyses focused on that topic, and the importance of this research. ○ Why the conceptual evolution is shown before the strategic diagrams? It does not make sense. Thus, Figure 8 (strategic diagrams) must be place before Figure 7 (evolution).

○
Last paragraph of page 6 does not provide nothing interesting, since it is just describing the overlapping figure in a textual way. Authors should make an effort to show the knowledge learned.

○
Regarding the strategic diagrams, authors just put them but any comment is added. This is the most important part of the papers, and there is any comment describing them. Also, the strategic diagrams are difficult to understand since they have a lot of themes. I recommend to set a higher threshold for the data and network in SciMAT.
○ Why authors make a complementary analysis of the last five years in five different periods? In this subanalysis, authors made the same errors than in the analysis of the other periods. That is, no comments (or just a few) describing the results.