Research Intelligence
The Diffusion Index
How do we know where a technology sits on the adoption curve? We measure what fraction of its papers appear in top-tier journals. Early technologies are novel enough that just using them earns a Nature paper. Mature technologies don't.
In 2008, 118 out of every 1,000 NGS papers appeared in Nature, Science, or Cell. Simply demonstrating that next-generation sequencing could be done at scale was publishable in the top journals. The technology was so new that the method itself was the finding.
By 2024, that number had fallen to under 1 per 1,000 — a 153× decline in 16 years. NGS is now infrastructure. Labs don’t publish in Nature for using a sequencer any more than they publish for using a centrifuge. The technology has completed its journey from innovation to commodity.
That journey maps precisely to the Rogers adoption curve. And it leaves a measurable fingerprint in the publication record that we can track for any technology.
2008
Peak year
117.7 per 1,000
153×
Decline peak→today
2008 → 2024
~0.8
Commoditized floor
papers per 1,000 since 2016
How many papers, and where they land
Annual NGS papers split into three tiers: Nature / Science / Cell, the rest of the top-tier specialist journals (Tier 1+2), and everything else. On a log scale all three stay visible; the widening gap between them is the diffusion story the next chart measures as a ratio.
Log scale — each line is a count, so all three tiers stay visible despite spanning four orders of magnitude. The gap between the lines is the dilution: top-3 output barely moves while total volume explodes. 2006–2025 (2026 partial year omitted).
NGS — journal prestige over time
Two lenses on the same decline, 2006–2026. Top-tier share: papers per 1,000 landing in top journals. Journal H-index: the median (or mean) Scimago H-index of the journals NGS papers appear in. Shaded regions are Rogers adoption phases. Based on 105,432 PubMed-indexed NGS papers.
Peak: 268.9 per 1,000 in 2008 · current (2025): 5.5 per 1,000 (Laggards era)
On the H-index view. Each paper inherits its journal’s Scimago H-index (a journal-prestige proxy, not per-paper citations); matched for 83% of papers by journal name. The median journal H-index fell from 284.5 (2008) to 123 (2025) — NGS moved off the elite journals into the mainstream. Median is the more honest line: Scimago H-index is cumulative, so big, long-running megajournals score high on volume alone.
Where NGS publishes, by H-index. Nature 1495 · Science 1433 · Cell 956 · PLOS ONE 500 · Scientific Reports 382 · Frontiers in Microbiology 287 · Methods in Molecular Biology 229. The early elite ceiling (Nature, Cell) is now rare; the bulk of NGS work lands in high-volume generalist and specialist journals (PLOS ONE, Scientific Reports, Frontiers, BMC Genomics).
What was hot in Nature, Science & Cell
Only ~175 NGS papers ever reached the top three journals across two decades. Classified by research theme, they show the frontier rotating every two to three years — and 59% are basic research, not clinical. That is the supply-side proof that the diffusion score measures discovery, not clinical value. Read the full analysis →
Absolute counts — each block is a real paper. Every NGS paper in Nature, Science, or Cell (2006–2025), by research theme — absolute counts, since the top-3 set is small (1–16 papers/year), so each block is a real paper. Themes classified by Claude (Haiku) from titles.
What the curve tells us
The 2008 peak was the “novelty premium.”
At 118 per 1,000, getting into Nature just required doing sequencing. Illumina had launched its first commercial short-read platform in 2006. The 2008 peak is the moment when early adopters realized the technology worked at scale — and every top lab rushed to show what they could do with it.
The chasm crossed in 2009; commoditization came around 2014–2016.
The Adoption Index dates the chasm — early adopters to early majority — to 2009, when the novelty premium first collapsed below a third of its peak. Full commoditization followed: the score fell from 5.4 in 2013 to 0.86 in 2016, a 6× decline, as NGS went from “exciting technology that smart labs are using” to “standard technique in any molecular biology lab.” Both are milestones of scientific adoption — clinical and commercial uptake ran years behind.
The floor at ~0.6 is the permanent residual.
From 2016 onwards, the score has stayed remarkably flat between 0.6 and 1.3. This is the “application novelty premium” — NGS itself is routine, but a truly novel application of NGS (a landmark GWAS, a new clinical use, a breakthrough in single-cell) can still earn a top journal. The floor doesn’t go to zero; it settles where the rate of genuinely novel applications stabilizes.
Using this to locate other technologies on the curve.
Spatial transcriptomics currently has approximately 8–15 per 1,000 papers in top-tier journals — which is where NGS was in 2009–2011. If the diffusion dynamics are similar, spatial transcriptomics has roughly 5–8 years before commoditization. Single-cell is likely a few years behind NGS on this curve. Long-read sequencing has crossed into early majority — adopted as infrastructure, though it never earned the same novelty premium. These are not predictions — they are calibrated estimates based on the NGS precedent.
Where the papers come from
Share of NGS papers by first-author affiliation, 2006–2025. Parsed from 105,432 affiliations in the database.
In 2006, two-thirds of all NGS papers came from US institutions. The technology was invented and commercialized in the US and UK — Illumina in San Diego, Solexa in Cambridge — and early access was concentrated accordingly.
China entered the top-three by 2012 at 10%, reached parity with the US around 2016, and overtook the US in 2018 — never looking back. At its 2021 peak, 42% of sampled NGS papers came from Chinese institutions. BGI Genomics, which built one of the world's largest sequencing operations in Shenzhen, is part of the explanation. So is China’s systematic investment in genomics infrastructure across its top research universities.
Germany has been a quiet constant throughout — typically 5–8%, reflecting the strength of the Max Planck institutes, EMBL, and the Helmholtz centers. India is beginning to appear from 2023 onward, following the same pattern China showed a decade earlier: infrastructure builds, then output follows.
What the sequencing was for
The application mix of NGS papers, 2006–2025, classified into research, clinical, and other uses. Normalized to 100% per year to show how the purpose of sequencing shifted as the technology diffused.
Research
Clinical
Other
Each bar is one year, normalized to 100% to show the mix (volume grew ~600× over the period — see the diffusion curve above). Other applications combines the former “applied markets” bucket — agriculture, environmental and food-safety sequencing, veterinary and microbiome work — with method-development papers and those that don’t fit neatly into one of the research or clinical categories. Classified by Claude (Haiku) over 98,417 abstract-bearing NGS papers, 2006–2025.
In 2006, NGS was almost entirely a research instrument: roughly three-fifths of papers were basic research or population studies, and clinical diagnostics were a rounding error. The early adoption curve was driven by labs asking what the technology could see, not clinicians using it to make decisions.
By 2025 the picture has inverted. Clinical applications — oncology profiling, NIPT, rare-disease and other diagnostics, and liquid biopsy — are now the largest block at roughly half of all NGS papers, while the research share has fallen by about two-thirds. This is the same transition the diffusion curve measures, seen from the demand side: a discovery platform becoming clinical infrastructure. The novelty premium collapsed because the work moved from the bench to the clinic — exactly where a maturing precision-medicine technology should end up.
Who uses it for what
Cross the application mix with first-author geography and the four big NGS regions turn out to have distinct fingerprints — and not the obvious ones. Each cell is an application’s share within that region; color shows how far it deviates from the average of the selected regions (green above, red below). Click any region to recompute the comparison around the set you care about — the figures below describe the default four-region view.
Click a region to add or remove it from the comparison. The Average column and the green/red deviations recompute for whatever set you pick — so you can, say, keep only two regions and see how they differ from each other.
| Application | Average 4 regions | |||||
|---|---|---|---|---|---|---|
| Research | ||||||
| Academia / basic research | 21.5% | 32.2% +49% | 20.7% -4% | 19.6% -9% | 17.0% -21% | 17.0% |
| Population studies | 3.8% | 5.8% +53% | 4.9% +28% | 3.2% -17% | 2.0% -49% | 5.0% |
| Biopharma R&D | 4.9% | 6.0% +23% | 5.1% +4% | 5.4% +11% | 4.1% -17% | 4.6% |
| Clinical | ||||||
| Clinical Dx — oncology | 13.8% | 17.6% +28% | 14.5% +6% | 23.1% +68% | 10.1% -27% | 16.2% |
| Clinical Dx — other | 20.7% | 13.5% -34% | 24.6% +19% | 18.8% -9% | 21.2% +3% | 22.5% |
| Clinical Dx — NIPT | 1.4% | 1.0% -30% | 1.3% -6% | 0.9% -35% | 1.7% +24% | 1.5% |
| Liquid biopsy — Dx | 2.2% | 2.2% -1% | 2.5% +11% | 3.2% +42% | 1.9% -14% | 2.1% |
| Biopharma clinical trials | 0.9% | 1.2% +30% | 1.1% +20% | 1.6% +73% | 0.5% -40% | 1.3% |
| Liquid biopsy — trials | 0.6% | 0.5% -14% | 0.7% +24% | 1.1% +91% | 0.4% -23% | 0.7% |
| Other | ||||||
| Epidemiology / ID | 5.5% | 6.4% +18% | 6.5% +20% | 5.3% -3% | 4.1% -25% | 8.0% |
| Other applications | 24.8% | 13.6% -45% | 18.1% -27% | 17.8% -28% | 37.0% +49% | 21.2% |
Each cell shows that application’s share within the region, and below it the deviation from the average of the selected regions (e.g. +49% = 1.49× that average). Greyed columns are excluded from the baseline. Cumulative over 102,631 classified papers with a parseable first-author affiliation; region assigned by keyword match. "Other applications" = applied markets, method development, and papers not fitting one of the named categories.
The United States is the research engine — not the clinical one. It over-indexes most heavily on population and basic research (about 1.5× the four-region average), and biopharma R&D, and under-indexes the applied and routine-diagnostic categories. The technology was invented there, and the US literature still reflects a field asking what sequencing can reveal.
Japan is the clinical-translational specialist: small in volume but the most over-indexed region on oncology diagnostics, biopharma clinical trials, and liquid biopsy — it uses NGS to run trials and treat cancer. China is the opposite outlier, over-indexing on applied and method-heavy work and on prenatal testing (NIPT) while under-indexing population genetics, clinical oncology, and trials. Europe is the balanced generalist, closest to the global average across almost every category.
The intuition that the US and China are both racing into clinical applications doesn’t survive contact with the data: the US leans research, China leans applied and prenatal, and it is Japan — a tenth of China’s volume — that punches above its weight in the clinic.
Methodology: 105,432 papers from PubMed matching “next generation sequencing” OR “next-generation sequencing” OR “high-throughput sequencing,” 2006–2026. Journal tiers: Tier 1 = Nature, Science, Cell (is_top3); extended Tier 1+2 includes Nature Methods, Nature Biotechnology, Nature Medicine, Nature Genetics, Nature Communications, Genome Biology, and other high-impact specialist journals (H-index ≥ 150). Diffusion score = top-tier papers / total papers × 1,000. Stored in Supabase, computed via the diffusion_index_by_year view. Geography: first-author affiliation parsed via city/country keyword matching; years ≥ 2012 sampled at 1,000 papers per year (REST API limit), years 2006–2011 complete. Shares are % of sampled papers for that year.