Overview of Trends in Global Single Cell Research Based on Bibliometric Analysis and LDA Model (2009–2019)

The data were drawn from Clarivate Analytics’ Web of Science (WoS) Core Collection in April 2020. We used the following search query: TS=((“single cell*”) NOT (“fuel-cell*” or “membrane-fuel-cell*” or “oxide-fuel-cell*” or “yuannhsolid-oxide-fuel-cell*” or “SOFC” or “proton-exchange-membrane-fuel-cell*” or “PEMFC” or “Direct-methanol-fuel-cell*” or “vanadium-redox-flow-battery*” or “solar-cell*” or “membrane-electrode-assembly” or “electrocatalyst” or “electrolyte” or “oxygen-reduction-reaction” or “reactive-oxygen-species” or “electrode” or “cathode” or “anode” or “electric-field” or “conductivity” or “durability” or “electrochemistry” or “electrochemical-performance” or “electrooxidation” or “Impedance” or “Impedance spectroscopy” or “Polymer-electrolyte-membrane” or “graphene” or “algorithm”)). The study was restricted to peer reviewed research papers (articles and reviews) between 2009 and 2019. Meeting abstracts, proceedings papers, notes, corrections, editorial material, and letters were thus excluded. A total of 30,804 publications were collected.

Thomson Data Analyzer (TDA) and Microsoft Excel were employed for bibliometric study. Gephi was utilized for national collaboration analysis. The size of nodes represents the number of publications and the size of the lines represents the frequency of collaboration.

According to the Perplexity-avg_sim curve and topic identification effect, K=20 was chosen as the optimum topic number (Figure 3). The top30 frequent terms with highest probability in the topic-keywords distributions were chosen to interpret the identified topics. After filtration of insignificant and repetitive topics, nineteen topics were identified. Due to the space limitation, only the top20 frequent terms of each topic are displayed in Table 1.

Figure 3

These topics can be divided into three categories. The first one is about single cell research methods, which include topic2, 7, 10, and 17. The second one is research on mechanism of biological process on single cell level, which include topic0, 1, 4, 5, 6, 8, 9, 11, 12, and 18. The third one is about single cell research application, which include topic3, 13, 14, 15, and 16.

Topic evolution trends were obtained by calculating topic strengths of each year and the results are listed in Figure 4. The x-coordinate represents the year, and the y-coordinate represents the topic strength. Strengths of some topics are on the rise. Topic3 “Immune response”, topic5 “Phylogeny on single cell level”, topic8 “Molecular mechanism of embryonic development”, topic12 “Cell-to-cell variability analysis” and topic16 “Cellular heterogeneity analysis” show increase tendency in fluctuation. The topic strengths of topic13 “Cancer diagnosis and treatment” is nearly continuously increasing, except for a brief dip in 2015. The strength of some topics are in general decline. Topic4 “Signal transduction”, topic6 “Intracellular calcium modulation”, topic17 “Single cell living imaging” show trends of fluctuating downward. The topic strength of topic7 “Single cell gel electrophoresis” continues to drop. The topic strength of topic0 “Pathology of brain disease”, topic1 “Mathematical modeling of cell cycle”, topic2 “Single cell detection platform”, topic9 “Cell adhesion”, topic10 “Isolation and sorting of single cell”, topic11 “Cell migration”, topic14 “Single cell oil”, topic15 “Stem cell” and topic18 “Biolfilm formation” show large volatilities. The strength of almost all the topics refers to the method category tends to decrease since 2015, while the topics refers to the clinical application of single cell technologies such as topic13, 15, and 16 show rapid growth since 2015. This indicates that technological innovation drives the development of applied research.

Figure 4

Research topic distribution of countries is usually concerned about in the scientific and technology information analysis. Traditional bibliometric or co-occurrence methods are either too tedious or unable to perform quantitative analysis. Taking the post-discretized method which is used for temporal distribution analysis for reference, the topics can also be dispersed to different countries to detect the spatial distribution. Thus, the research topic distribution can be easily observed through the topic strength analysis. As is shown in Figure 2a, the top10 most productive countries of single cell research field are the US, China, Germany, UK, Japan, France, Canada, Switzerland, Italy, and the Netherlands (from high to low). The topic identification results were dispersed into these ten countries and the topic strengths of each country were calculated. Comparing the topic strengths of every topic in each country, the information about the research investment emphasis and development priorities can be obtained (Figure 5).

Figure 5

The strengths of research topics reflects the country's focus on different topics or research and development investments. For example, China and Japan paid far more attention to topic2 “Single cell detection platform” than other topics. As for Italy, the topic that attracted the most attention was topic7 “Single cell gel electrophoresis”. The topic distribution of Germany and the Netherlands are relatively balanced.

From the two dimensions of time and space, the topic distribution trends of countries can be analyzed. Taking China for example, the topic distribution trend from 2009 to 2019 can be observed in Figure 6. For the convenience of comparison, the topic strength range for each year as 0.04–0.07, with an interval of 0.005. From 2009 to 2013, topic7 “single cell gel electrophoresis” was the topic with the highest strength. Since 2014, the superiority of topic7 had been replaced by topic2 “single cell detection platform”, and the standard deviation of topic strengths is significantly lower than which in previous years.

Figure 6

In terms of a specific topic, the evolution trends of each country can be analyzed. Taking topic13 “Cancer diagnosis and treatment” as an example, the evolution trends of each country can be observed from Figure 7. For the convenience of comparison, the topic strength range for each year as 0.042–0.056, with an interval of 0.002. For China, France, and Switzerland, the topic strength increase dramatically since 2015, for Germany, UK, Japan, Canada, and Italy, the topic strength are fluctuate during the last decade.

Figure 7