Spatial-Temporal Evolution of Scientiﬁc Production about Genetically Modiﬁed Maize

: Maize is the grain cereal that is the basis of human and animal diets in Mexico and Latin America; it constitutes an essential crop for global food security. The objective of this study was to analyze the spatial–temporal evolution of scientiﬁc production on the theme of GMO maize, through a bibliometric analysis of the texts available in the main editorial houses (Elsevier, Scopus, and Springer), open access journal articles database (Conricyt, Scielo, Redalyc, Latindex, Claryvate Analytics, Periodica, and DOAJ), and freely accessible web search engine Google Scholar, to determine the factors that inﬂuence the impact of the studies. From 1991 to 2019, 917 texts were found whose spatial–temporal evolution showed a linear growth that concentrated in Latin America (58.56%). The low impact (measured by the number of bibliographic citations) of scientiﬁc studies developed in countries of Latin America was related to their publication in journals edited in their own countries and in Spanish, which restricts the constructive criticism of peer review. For the case of Mexico, a spatial discrepancy was also found between research centers and production areas, which limits the transference of technology; and no specialized author in theme of GMO maize was found; the researchers responded to “scientiﬁc trends” in agreement with the agrarian policies of the time.


Introduction
Maize (Zea mays L.) is the cereal with highest worldwide production level, above wheat (Triticum aestivum L.) and rice (Oryza sativa L.), and it constitutes the basis of the human diet for countries of Latin America, especially Mexico, where it is considered a strategic crop for the country's food security [1]. With the development of biotechnology, GMO maize varieties have been created with resistance to insects and herbicides, which allows productivity to increase by area [2].
Maize is originally from Mexico, territory where 64 races are cultivated derived from the millenary domestication process of its wild relatives, Teocintle and Tripsacum, in combination with different environments, agricultural systems and ethnic groups [3]. However, despite this diversity of maize races, Mexico is not the main maize producer in the world. Of global production, 70.06% is concentrated in five countries, with USA (35.42%) and China (21.93%) as the leading producers, followed by Brazil (7.22%), Argentina (3.01%), and Mexico (2.48%) [1].
Unlike Mexico, USA, China, Brazil, and Argentina base their productivity on the use of GMO maize varieties, which provide higher yields in the field than the conventional varieties [4]. Reference [5] found that the agricultural competitiveness of a country is directly related with the quality of the research performed in this sector. To measure the quality of research, [6] suggest a methodology of bibliometric analysis that allows studying and analyzing the evolution of scientific activity through its publications.
The publication of a scientific study is the most effective way to transmit knowledge acquired as consequence of the research, and its visibility is important for the researchers themselves, for the institutions where they work, and for the organizations that finance the research [7]. The growth of scientific production in recent decades and their indexation in automatized bibliographic databases have strengthened the use of bibliometry and the generation of indicators to measure the results of scientific and technological activity [8].
Bibliometric studies based on the scientific articles published are supported by the application of quantitative methods, generation of indicators and mathematical models, which allow characterizing their development and evolution [9]. Through bibliometric indicators, the journals and other periodic publications can be evaluated objectively, and the trends that they follow can be observed to generate useful information to improve their management [7].
Understanding the scientific information that is developed around a topic allows making decisions in relation to its improvement [10]. Bibliometric studies have been developed in the agricultural sector to evaluate general themes of agronomy [6], specific themes such as ecological agriculture [11], fruits and vegetables [12], and even for agronomists in particular [13], agricultural journals [14], national agricultural sectors [8], and specific crops such as wheat and barley [15], and rice [9,16].
However, despite the growth in the number of publications that has taken place in the agricultural sector and specifically in themes related to transgenic maize, there is scarce research to assess the evolution and the impact of these publications at the level of the user [17]. Considering this, the objective of the present study was to analyze the spatial-temporal evolution of scientific production on the theme of GMO maize at the global level, and with emphasis for Mexico, through text mining (bibliometry) and Social Network Analysis (SNA) to determine the factors that impact the quality of the studies.

Origin of the Information and Data Preparation
Scientific articles about transgenic maize available in the principal editorial houses (Elsevier, Scopus, and Springer), open access journal articles database (Conricyt, Scielo, Redalyc, Latindex, Claryvate Analytics, Periodica, and DOAJ), and freely accessible web search engine Google Scholar were considered in this study. They were compiled from January to April 2020. It was considered all scientific articles available up 2019. The keywords used in the search were "maíz transgénico", "maíz genéticamente modificado", "genetically modified corn" and "transgenic corn", identifying them in the titles and keywords of the publications. In addition, the "snowball" technique was used to obtain the missing articles, from the reference list of the articles initially found. However, it has to be considered that the technique of "snowball" as being non-probabilistic, has the possibility of incurring in bias, while recovering the scientific articles, favoring the documents in English. For upcoming this bias, in the search of the scientific articles, the keywords used were both in English and Spanish, what let to obtain the great majority of important publications [18].
In order to establish the bibliometric indicators, the variables analyzed from each of the articles were "name of the journal, editing institution, country of edition", and "language of publication", which served to determine the profile of the journals that publish studies related to the theme of "GMO maize; first author and collaborators" which served to understand the network of actors involved in the research; "year" used in order to place the Agriculture 2021, 11, 246 3 of 14 information in a time line; "institution of the first author" and "the first author's country of origin" to evaluate the frequency of publications of institutions by country.
The "postal code of the institution of the first author" served to perform a geographic localization of the information's institution of origin; in the cases when the postal address did not appear, the name of the institution was searched with Google Earth ® tools and in the official pages of the institutions. The "title, abstract" and "keywords" were used to categorize the topic addressed by the publication in agreement with the classification by the National Resource Consortium of Scientific and Technological Information [19] for genetically modified maize. Finally, the usefulness of the publications was determined with the "number of citations".
Capturing the variables was done in a worksheet, in which they were systematized and classified in thematic areas according to their content; in addition, the original language of each of the texts was respected. While capturing all the information, some records were standardized because the information available in the articles was sometimes incomplete or presented with variations [20]. In addition, special characters were eliminated or changed to ease the analysis, such as: ñ (by n), accent marks, superindex, subindex, ® , ©, among others. Due to interpretation reasons, the information for the case of Mexico was separated into three periods that coincide with the periods of three government administrations: [2001][2002][2003][2004][2005][2006][2007][2008][2009][2010][2011][2012], and 2013-2018.

Analysis with Text Mining
With the help of the RcmdrPlugin.temis complement of the statistical software R [21], the number of articles per year and journal, the institution where the research originated, and the articles where the main author is from the same institution of the journal were obtained.

Network Analysis
With the Sci2tool software [22] the interactions present between the first authors and collaborators were analyzed with the aim of understanding the consistency of the researcher's work; that is, to assess whether the author has published just in one year or else has been publishing constantly throughout time, which gives an idea of the author's consolidation on the topic of GMO maize. The syntaxis used with the Sci2tool software was Extract bipartite Network, and for its visualization the Gephi software was used [23].
Finally, with the help of the variable "postal code" and Google Earth ® tools, the geographic coordinates in tenths of degree were obtained (Longitude, Latitude) of each of the articles published. The spatial representation of the number of articles per institution was carried out with the geographic package ARGIS ® [24].

Scientific Production on GMO Maize in the International Context
From 1991 to 2019, a total of 917 scientific articles were published, which resulted in 20,997 bibliographic citations (Figure 1). The first study recorded dates from 1991; however, since 2000 a growing production was seen for the topic of "GMO maize". The period of highest productivity was from 2008 to 2018 with 76.  49), and China (4.58%, 42). The fact that the countries with highest scientific production are the countries with greatest maize grain production globally stands out ( Figure 2). Figure 2 also shows that the countries with highest number of publications are located in Latin America, which explains that 55.29% (507 articles) are  42). The fact that the countries with highest scientific production are the countries with greatest maize grain production globally stands out ( Figure 2). Figure 2 also shows that the countries with highest number of publications are located in Latin America, which explains that 55.29% (507 articles) are published in Spanish, 39.59% (363) in English, 3.93% (36) in Portuguese, and 1.20% (11) in other languages (primarily French, Mandarin, and Russian).      In the case of the main countries that recorded scientific production about GMO maize, the themes of productivity (26.28%, 241 articles), pests and diseases (12.65%, 116), and com-Agriculture 2021, 11, 246 5 of 14 petitiveness (12.10%, 111) had the highest frequencies in scientific articles ( Figure 3). [25] found that the policy that the USA adopted to encourage the development of technological innovations, allowed it to become the principal maize producer in the world. In contrast with Mexico, where the development of research consisted mostly in demonstrating the irrelevance of transgenic maize [17,25].
In the case of the main countries that recorded scientific production about GMO maize, the themes of productivity (26.28%, 241 articles), pests and diseases (12.65%, 116), and competitiveness (12.10%, 111) had the highest frequencies in scientific articles ( Figure  3). [25] found that the policy that the USA adopted to encourage the development of technological innovations, allowed it to become the principal maize producer in the world. In contrast with Mexico, where the development of research consisted mostly in demonstrating the irrelevance of transgenic maize [17,25]. In the 917 articles analyzed, 703 first authors were found, between author and collaborators they were 2360 different individuals in total. The mean of co-authorship was 2.57 and the mode (160) of 3 authors per article, with extreme values of a single author (29 articles) and 6 studies with more than 12 authors. The network of authors and collaborators ( Figure 4) was made up of 2360 nodes (authors) and 2132 aspects (links). The links in a SNA are important because it is through them that an author can reach certain ideas, knowledge and information that is socially distant for him [26].
The density of the network had a value of 0.001, implying that for the topic of GMO maize there is not much collaboration between authors in the international context. This is evident since there are 501 institutions (from 917 studies) indicated as the ascription of the first author. The density is an indicator of the SNA that implies that both the nodes interact (are linked) between one another, mathematically it is a value within the interval (0 to 1), the closer to 1 the higher the interaction in the network [27].
A subnetwork formed by Mexican researchers was detected ( Figure 4): Espinosa_Cal-deron_A, Mejia_Andrade_H and Turren_Fernandez_A from INIFAP; Cortes_Flores_JI (UNAM); Serratos_Hernandez_JA (UACM) and Massieu_Trigo_Y (UAM). As a whole, the In the 917 articles analyzed, 703 first authors were found, between author and collaborators they were 2360 different individuals in total. The mean of co-authorship was 2.57 and the mode (160) of 3 authors per article, with extreme values of a single author (29 articles) and 6 studies with more than 12 authors. The network of authors and collaborators ( Figure 4) was made up of 2360 nodes (authors) and 2132 aspects (links). The links in a SNA are important because it is through them that an author can reach certain ideas, knowledge and information that is socially distant for him [26].

Bibliometric Indicators
The 917 studies analyzed were published in 449 scientific journals. A total of 248 articles, 27.05% and 8628 bibliographic citations, 41.09%, were found in 20 journals (Table  1). Among these 20 main journals, 13 are from Latin American countries: Mexico (5), Colombia (4), Argentina (1), Cuba (1), Venezuela (1), and Brazil (1); they publish mostly in The density of the network had a value of 0.001, implying that for the topic of GMO maize there is not much collaboration between authors in the international context. This is evident since there are 501 institutions (from 917 studies) indicated as the ascription of the first author. The density is an indicator of the SNA that implies that both the nodes interact (are linked) between one another, mathematically it is a value within the interval (0 to 1), the closer to 1 the higher the interaction in the network [27]. The

Bibliometric Indicators
The 917 studies analyzed were published in 449 scientific journals. A total of 248 articles, 27.05% and 8628 bibliographic citations, 41.09%, were found in 20 journals (Table 1). Among these 20 main journals, 13 are from Latin American countries: Mexico (5), Colombia (4), Argentina (1), Cuba (1), Venezuela (1), and Brazil (1); they publish mostly in Spanish and do not have a JCR impact factor or a very low factor (Q4), which according to [29] limits the number of bibliographic citations since English is the language adopted as universal by the scientific community.
The low impact (bibliographic citations) of the scientific studies developed in countries of Latin America is related to the fact that the researchers publish their studies in national journals (Table 2), edited in Spanish; meanwhile, the journals of highest impact are found in English-speaking countries (United Kingdom and USA) [31], published in English. This explains the high impact of articles from USA, with 63.03% of their studies published in journals of their own country, and the remaining percentage in journals of other countries also edited in English.
Among the 20 most cited studies about GMO maize themes, 14 belong to a first author whose institution of ascription is localized in USA and only one study corresponds to a researcher in Latin America (Mexico); they have all been published in English and in journals of English-speaking origin (Table 3). According to [5] the economic development of a country is directly related to the quality of the research it performs; countries with consolidated economies (USA, Switzerland, Italy, and China) invest more in their research centers, which allows them higher technological development, compared to underdeveloped economies like the Latin American where investment in research is lower. In addition, some of these consolidated economies (Switzerland, Italy, and China), whose native language is not English, also publish their studies of higher impact in this language, the same in which most of the indexed journals included in the Journal Citation Report are edited, which is included in the Web of Science (WoS) and belongs to an information company in USA, Thomson Reuters [30].

Scientific Production on the Theme of GMO Maize in Mexico
From 1991 to 2019 Mexican researchers published 225 scientific articles on the topic of GMO maize. From the 225 studies, 163 (72.44%) were developed in 11 of 50 institutions, taking as reference the institution of ascription of the first author. The institutions with highest productivity were UAM (36 studies), UNAM (34), COLPOS (19), CHAPINGO (15), INIFAP (14), and IPN (12). Due to tradition, in CHAPINGO (founded in 1854), COLPOS (1959) and INIFAP (1985), the studies published stemmed from the initiatives to develop agricultural research.
The spatial distribution of the institutions with scientific productivity on GMO maize (1991-2019) ( Figure 5) allows deducing that research on this topic is located in the center of the country, while in terms of the states with highest maize grain production in Mexico from 2000 to 2019, the areas with highest production are in the north (Sinaloa and Chihuahua), east (Jalisco, Michoacán, Guanajuato, and Estado de México), northeast (Tamaulipas and Veracruz), and southeast (Chiapas) of Mexico. This aspect of the centralization of research has been addressed by [10], who found that the discrepancy between the production areas and the research centers makes the transference of technology difficult.

Scientific Production on the Theme of GMO Maize in Mexico
From 1991 to 2019 Mexican researchers published 225 scientific articles on the topic of GMO maize. From the 225 studies, 163 (72.44%) were developed in 11 of 50 institutions, taking as reference the institution of ascription of the first author. The institutions with highest productivity were UAM (36 studies), UNAM (34), COLPOS (19), CHAPINGO (15), INIFAP (14), and IPN (12). Due to tradition, in CHAPINGO (founded in 1854), COL-POS (1959) and INIFAP (1985), the studies published stemmed from the initiatives to develop agricultural research.
The spatial distribution of the institutions with scientific productivity on GMO maize (1991-2019) ( Figure 5) allows deducing that research on this topic is located in the center of the country, while in terms of the states with highest maize grain production in Mexico from 2000 to 2019, the areas with highest production are in the north (Sinaloa and Chihuahua), east (Jalisco, Michoacán, Guanajuato, and Estado de México), northeast (Tamaulipas and Veracruz), and southeast (Chiapas) of Mexico. This aspect of the centralization of research has been addressed by [10], who found that the discrepancy between the production areas and the research centers makes the transference of technology difficult. The language preferred by Mexican researchers to publish their results on GMO maize was Spanish with 88.44% (199) of the studies, followed by English for the remainder 11.56% (26). The themes of most interest were those related with competitiveness (56 studies, 24.89%) and productivity (49, 21.78%) directed at showing the irrelevance of GMO maize in the Mexican farmland. The 225 studies were published in 123 different journals, 86 of which were edited by national institutions that in total published 177 articles (78.67%).
In Mexico, most of the researchers published their studies about GMO maize in journals edited by their institution of ascription (Table 4), which restricts the constructive criticism of peer review and, according to [51,52], is indicator of the low impact of the publications, since it limits the feedback from interdisciplinary groups regarding the relevance of the research. In this regard, [10,53] found that research in Mexico not only presents "academic endogamy" but also that the researchers in their eagerness to remain in the Sistema Nacional de Investigadores-Consejo Nacional de Ciencia y Tecnologia (SNI-CONACYT) tend to replicate the same methodologies in different study areas, which limits innovation in research and allows the circularity (redundancy in the study object) of the publications. The language preferred by Mexican researchers to publish their results on GMO maize was Spanish with 88.44% (199) of the studies, followed by English for the remainder 11.56% (26). The themes of most interest were those related with competitiveness (56 studies, 24.89%) and productivity (49, 21.78%) directed at showing the irrelevance of GMO maize in the Mexican farmland. The 225 studies were published in 123 different journals, 86 of which were edited by national institutions that in total published 177 articles (78.67%).
In Mexico, most of the researchers published their studies about GMO maize in journals edited by their institution of ascription (Table 4), which restricts the constructive criticism of peer review and, according to [51,52], is indicator of the low impact of the publications, since it limits the feedback from interdisciplinary groups regarding the relevance of the research. In this regard, [10,53] found that research in Mexico not only presents "academic endogamy" but also that the researchers in their eagerness to remain in the Sistema Nacional de Investigadores-Consejo Nacional de Ciencia y Tecnologia (SNI-CONACYT) tend to replicate the same methodologies in different study areas, which limits innovation in research and allows the circularity (redundancy in the study object) of the publications. Finally, the network of authors and collaborators for the case of the publications where the first author was ascribed to a Mexican institution was divided into three periods: 2001-2006, 2007-2012, and 2013-2018. The periods proposed correspond to the policies that prevailed in the country during the different governments.
By  Table S1). According to [10] the increase in authors and articles that Mexico has been having in the agriforest sector is because the SNI-CONACyT awards more the quantity than the quality. In turn, [54] found that the agricultural policies adopted by Mexico during its different governments have encouraged the development of research on the theme of GMO maize.
With the promulgation of the Sustainable Rural Development Law in 2001 [55], research began in Mexico on the topic of GMO maize, while before that the studies were centered on attaining the best yields with native germplasm or improved through hybridization [54]. The government in 2007-2012 encouraged the use of GMO seed and the studies were focused on highlighting the geopolitical irrelevance of this biotechnology in Mexico as center of origin and domestication of maize [17]. Finally, the government in 2013-2018 motivated the use of improved maize varieties (native) through the Traditional Agriculture Sustainable Modernization Program (MasAgro) and the studies revolved around arguing that it is possible to achieve self-sufficiency (food security) without the need of resorting to the use of GMO varieties [54]. Of the scientific articles analyzed, 90% agree that it is not convenient the growth of GMO maize in Mexico. The main argument to this position is that Mexico is the main center of origin and domestication of maize, so there is a great genetic diversity of this species and there are close cultural connections between maize and the different native peoples.
The co-authorship networks built according to the degree of entry of the authors in the network allowed identifying the importance (reference) of the author in the network, by quantifying the number of links that an author receives from others [27]. In 2001-2006, there was not an author with high reference, and three authors were found with medium reference (Soberon_J, Schoel_B and Ezcurra_E) ( Figure S1). In 2007-2012, there was an author with high reference (Castillo_González_F) (Figure S2), and in 2013-2018, two (Vizacarra_Bordi_I and Avila_Castaneda_JF) (Figure 6), the authors of medium reference increased in 5 and 18, respectively (Table S1). However, despite these increases, from 2001 to 2018 there was not an author observed whose research was constant, which points to there not being in the country any specialists in the theme, but rather research responded to scientific trends in agreement with the agrarian policies of the moment.  Table S1). According to [10] the increase in authors and articles that Mexico has been having in the agriforest sector is because the SNI-CONACyT awards more the quantity than the quality. In turn, [54] found that the agricultural policies adopted by Mexico during its different governments have encouraged the development of research on the theme of GMO maize.
With the promulgation of the Sustainable Rural Development Law in 2001 [55], research began in Mexico on the topic of GMO maize, while before that the studies were centered on attaining the best yields with native germplasm or improved through hybridization [54]. The government in 2007-2012 encouraged the use of GMO seed and the studies were focused on highlighting the geopolitical irrelevance of this biotechnology in Mexico as center of origin and domestication of maize [17]. Finally, the government in 2013-2018 motivated the use of improved maize varieties (native) through the Traditional Agriculture Sustainable Modernization Program (MasAgro) and the studies revolved around arguing that it is possible to achieve self-sufficiency (food security) without the need of resorting to the use of GMO varieties [54]. Of the scientific articles analyzed, 90% agree that it is not convenient the growth of GMO maize in Mexico. The main argument to this position is that Mexico is the main center of origin and domestication of maize, so there is a great genetic diversity of this species and there are close cultural connections between maize and the different native peoples.
The co-authorship networks built according to the degree of entry of the authors in the network allowed identifying the importance (reference) of the author in the network, by quantifying the number of links that an author receives from others [27]. In 2001-2006, there was not an author with high reference, and three authors were found with medium reference (Soberon_J, Schoel_B and Ezcurra_E) ( Figure S1). In 2007-2012, there was an author with high reference (Castillo_González_F) (Figure S2), and in 2013-2018, two (Vizacarra_Bordi_I and Avila_Castaneda_JF) (Figure 6), the authors of medium reference increased in 5 and 18, respectively (Table S1). However, despite these increases, from 2001 to 2018 there was not an author observed whose research was constant, which points to there not being in the country any specialists in the theme, but rather research responded to scientific trends in agreement with the agrarian policies of the moment.

Conclusions
The spatial-temporal evolution of scientific production showed a linear growth of scientific studies worldwide about GMO maize, which were concentrated primarily in countries of America, with special emphasis in Latin America: Mexico, Colombia, Argentina, Brazil, and Cuba. The themes of greatest relevance were those related to productivity (26.28%), pests and diseases (12.65%), and competitiveness (12.10%). However, the relevance (measured by number of citations) of scientific production in Latin America was nearly null, as a result of the publication in journals edited in the countries themselves, in Spanish, when the journals of impact are led by English-speaking countries, in English. For the case of Mexico, a centralization of research was found with spatial discrepancy from the production areas; a phenomenon called academic endogamy which consists in publishing in journals edited by the authors' own institution; and an author with constant production was not observed during the period of analysis, but rather the researchers responded to "scientific trends" in agreement to the agrarian policies of the moment. Therefore, research on GMO maize in Latin America has a broad margin of improvement through the publication of texts in English and higher impact journals, which can contribute to agricultural competitiveness of the sector in each country.  Data Availability Statement: Availability of data. The data are available with the author for correspondence, at reasonable request.