38 Case Study 1: CRISPR Field Review (2010–2024)

R. Heller · Date: 2026-05-09 · Last changed: 2026-10-30

38.1 Objective

Map the structure, growth, and evolution of CRISPR gene editing research from its emergence to its current state using open bibliometric data.

38.2 Setup

library(tidyverse)
library(openalexR)
library(igraph)
library(tidygraph)
library(ggraph)
library(quanteda)
library(tidytext)
library(glue)
library(gt)

set.seed(20260509)

source(here::here("R", "api_helpers.R"))
source(here::here("R", "utils.R"))
source(here::here("R", "sci_palette.R"))

38.3 Data acquisition

works <- fetch_openalex(
  entity = "works",
  search = "CRISPR",
  from_publication_date = "2010-01-01",
  to_publication_date = "2024-06-30",
  type = "article",
  options = list(sample = 500, seed = 42)
)

works <- works |>
  mutate(year = year(publication_date))

cat(glue("CRISPR articles retrieved: {nrow(works)}\n"))
#> CRISPR articles retrieved: 500
cat(glue("Year range: {min(works$year)}--{max(works$year)}\n"))
#> Year range: 2010--2024

38.4 Publication growth

works |>
  count(year) |>
  ggplot(aes(x = year, y = n)) +
  geom_col(fill = palette_sci(1)) +
  labs(x = "Year", y = "Publications") +
  theme_sci()
Bar chart showing exponential growth in CRISPR publications from 2010 to 2024.

Figure 38.1: Annual publication output in the CRISPR field.

38.5 Citation landscape

ggplot(works, aes(x = cited_by_count)) +
  geom_histogram(binwidth = 10, fill = palette_sci(1), colour = "white") +
  labs(x = "Citations", y = "Papers") +
  theme_sci()
Histogram showing the heavily right-skewed citation distribution typical of a rapidly growing field.

Figure 38.2: Citation distribution of CRISPR articles.

works |>
  arrange(desc(cited_by_count)) |>
  head(10) |>
  select(display_name, year, cited_by_count, source_display_name) |>
  gt()
display_name year cited_by_count source_display_name
CRISPR-Cas12a-assisted nucleic acid detection 2018 1670 Cell Discovery
Prioritization of cancer therapeutic targets using CRISPR–Cas9 screens 2019 1452 Nature
Self versus non-self discrimination during CRISPR RNA-directed immunity 2010 715 Nature
CRISPR/Cas9 systems targeting β-globin and CCR5 genes have substantial off-target activity 2013 682 Nucleic Acids Research
Optimized gene editing technology for Drosophila melanogaster using germ line-specific Cas9 2013 427 Proceedings of the National Academy of Sciences
Combinatorial CRISPR–Cas9 screens for de novo mapping of genetic interactions 2017 392 Nature Methods
Small extracellular vesicles containing arginase-1 suppress T-cell responses and promote tumor growth in ovarian carcinoma 2019 337 Nature Communications
Two Distant Catalytic Sites Are Responsible for C2c2 RNase Activities 2017 336 Cell
Aneuploidy renders cancer cells vulnerable to mitotic checkpoint inhibition 2021 303 Nature
Coronavirus endoribonuclease targets viral polyuridine sequences to evade activating host sensors 2020 301 Proceedings of the National Academy of Sciences

38.6 Co-authorship network

author_data <- works |>
  select(id, authorships) |>
  unnest(authorships, names_sep = "_") |>
  select(work_id = id, author_id = authorships_id,
         author_name = authorships_display_name) |>
  filter(!is.na(author_id))

edges <- author_data |>
  inner_join(author_data, by = "work_id", suffix = c("_1", "_2"),
             relationship = "many-to-many") |>
  filter(author_id_1 < author_id_2) |>
  count(author_id_1, author_id_2, name = "weight")

g <- graph_from_data_frame(
  edges |> select(author_id_1, author_id_2, weight),
  directed = FALSE
) |> simplify(edge.attr.comb = list(weight = "sum"))

comp <- components(g)
giant <- induced_subgraph(g, which(comp$membership == which.max(comp$csize)))
V(giant)$community <- as.factor(membership(
  cluster_leiden(giant, resolution_parameter = 1.0,
                 objective_function = "modularity")
))
V(giant)$degree <- degree(giant)

cat(glue("Network: {vcount(giant)} nodes, {ecount(giant)} edges\n"))
#> Network: 46 nodes, 1035 edges
set.seed(42)
ggraph(as_tbl_graph(giant), layout = "fr") +
  geom_edge_link(alpha = 0.1, colour = "grey60") +
  geom_node_point(aes(size = degree, colour = community), alpha = 0.7) +
  scale_size_continuous(range = c(0.5, 5), guide = "none") +
  scale_colour_manual(values = palette_sci(n_distinct(V(giant)$community))) +
  labs(colour = "Community") +
  theme_void(base_family = "sans", base_size = 11) + theme(legend.position = "bottom")
Network graph showing collaboration patterns among CRISPR researchers, with communities coloured by Leiden algorithm.

Figure 38.3: Co-authorship network of CRISPR researchers.

38.7 Topic evolution

text_df <- works |>
  filter(!is.na(abstract), nchar(abstract) > 50) |>
  transmute(doc_id = id, text = paste(display_name, abstract, sep = ". "), year)

corp <- corpus(text_df, docid_field = "doc_id", text_field = "text")
toks <- tokens(corp, remove_punct = TRUE, remove_numbers = TRUE) |>
  tokens_tolower() |>
  tokens_remove(stopwords("en")) |>
  tokens_remove(c("study", "paper", "results", "using", "based"))

dfmat <- dfm(toks) |> dfm_trim(min_termfreq = 5, min_docfreq = 3)

top_by_year <- map_dfr(unique(text_df$year), function(yr) {
  docs <- docvars(dfmat, "year") == yr
  if (sum(docs) < 5) return(tibble())
  top <- topfeatures(dfmat[docs, ], 5)
  tibble(year = yr, term = names(top), freq = unname(top))
})
top_by_year |>
  group_by(year) |>
  mutate(term = reorder_within(term, freq, year)) |>
  ggplot(aes(x = freq, y = term)) +
  geom_col(fill = palette_sci(1)) +
  facet_wrap(~ year, scales = "free_y", ncol = 4) +
  scale_y_reordered() +
  labs(x = "Frequency", y = NULL) +
  theme_sci(base_size = 8)
Faceted bar chart showing the top 5 terms in each year of CRISPR research.

Figure 38.4: Top terms by year showing topical evolution.

38.8 Key findings

  1. Explosive growth: CRISPR publications grew exponentially from 2012, reflecting the rapid adoption of Cas9-based editing.
  2. Citation concentration: A small number of foundational papers dominate the citation landscape.
  3. Collaborative structure: The co-authorship network shows distinct communities, likely corresponding to different application domains (therapeutics, agriculture, basic biology).
  4. Topic evolution: Early terms focus on methodology; later years shift toward applications and clinical translation.

38.9 Lessons learned

  • OpenAlex sampling provides a representative snapshot but may miss some highly specialised or non-English publications.
  • The citation distribution is extreme: median citations are far below the mean, making median-based statistics essential.
  • Co-authorship networks in fast-growing fields are fragmented; many research groups work independently.
This book was built by the bookdown R package.