Collecting the total number of articles for each human gene (19504 human genes)

Load the following packages

library(rvest)
library(dplyr)
library(readr)

Read csv file

humangenes <- read_csv("/Users/Desktop/human_proteins.csv")
View(humangenes)
colnames(humangenes)

Select the genes and make them unique

humangenes1 <- humangenes %>%
  dplyr::select(Locus)%>%
  unique()
allgenes <- humangenes1$Locus
pubmed_data <- data.frame()

Start downloading the number of articles per human gene. This process depends on the speed of the internet.

for(i in allgenes){
  link <- paste0("https://pubmed.ncbi.nlm.nih.gov/?term=", i,"&sort=date")
  page <- read_html(link)
  gene_name <- i 
  number <- page %>% html_nodes("span.value") %>% html_text() %>% 
    replace(!nzchar(number), NA)
  pubmed_data<- rbind(pubmed_data, data.frame(gene_name, number, stringsAsFactors = FALSE))
}

Save the data

write.csv(pubmed_data1, "/Users/Desktop/pubmed_data06.04.2022", row.names = FALSE)

Load the following packages

library(cowplot)
library(tidyverse)
library(ggrepel)

Generate the plot

p <- ggplot(pubmed_genes, aes(x= fct_reorder(gene_name, number), y =number),
            alpha = 0.02 )+
  geom_point() +
  ylim(0, 30000)+
  geom_point(color = ifelse(pubmed_genes$number> 25000, "red", "black"))+
  theme_cowplot(font_size = 12)

Label the X and Y axis

p1 <- p + 
  theme(axis.text.x=element_blank(),
        axis.ticks.x = element_blank())+
  labs(x = "Human Genes", y = "Number of Articles", 
       title = "Number of Articles for each human gene")

Mark gene names with different color and label them

p2 <- p1 + geom_text_repel(aes(label=ifelse((pubmed_genes$number> 25000), gene_name, "")))

Create the interactive plot

library(plotly)
library(htmlwidgets)


p3 <- p2  %>%
  ggplotly(tooltip = c("x", "y"))
#save the plot
htmlwidgets::saveWidget(as_widget(p3), "index.html")