These are the most-cited research papers of all time

What is the most highly cited paper in the scientific literature? Chances are slim that most researchers know the answer. It turns out to be a 1951 publication in the Journal of Biological Chemistry describing an assay to determine the amount of protein in a solution¹. That has been cited more than 350,000 times in the Web of Science (WoS), a database that includes 98 million papers dating back to 1900.

In fact, many papers on biological laboratory techniques dominate the list of the most highly cited papers of all time, according to data provided to Nature by the US firm Clarivate, which owns the WoS. The top-100 list also includes papers on artificial intelligence (AI), research software and statistical methods.

The most-cited papers of the twenty-first century

The current list is an update of one that Nature published in 2014, which was topped by tools used in cell and molecular biology from the 1950s and 1970s (see Nature 514, 550–553; 2014). About half of the papers in that top 100 have now changed as citations to newer papers have soared. In 2014, it took about 12,000 citations to make the top-100 list; now, more than 30,000 citations are required. (This refers to citations from papers in the WoS ‘Core Collection’, used here to allow comparison with 2014.)

Nature’s news team also analysed records from two large research databases — Dimensions and OpenAlex — because their public-facing versions allow analyses back to 1900. These give slightly different rankings and greater citation counts, but generally feature similar papers. (Full details of the top-100 lists, including an analysis of median rankings across three databases, are in the Supplementary information.)

All of the lists are still headed by the 1951 paper. But a paper² by Microsoft researchers from a 2015 conference on AI is already ranked fifth, when median rankings are analysed across three databases (or seventh in the WoS alone; see ‘Top ten cited papers’).

Overall, 16 papers from the twenty-first century now make it into the all-time top 50, although they have had much less time than older papers to garner citations. Software descriptions and computer-aided findings have made notable rises in the ranking. (Bibliometricians point out that citations should be corrected for a paper’s year of publication and research field to make fair comparisons between disparate works.)

Rising annual volumes of research papers — meaning more references each year — and the greater visibility of that research online and on social media might explain how some modern papers have shot up the charts, says Paul Wouters, a retired scientometrics researcher at Leiden University in the Netherlands.

Science’s golden oldies: the decades-old research papers still heavily cited today

Another surprising trend is that some works from the 1980s and 1990s in density functional theory (DFT), an approximation that makes it easier to model the behaviour of electrons in materials, are gaining citations at an accelerating rate. They are among the most frequently referenced works in recent papers: three DFT articles are now in the all-time top ten. (For more on this trend, see ‘Science’s golden oldies: the decades-old research papers still heavily cited today’.)

The late US biochemist Oliver Lowry, who was lead author of the top-cited 1951 paper, was puzzled by its success. “Although I really know it is not a great paper … I secretly get a kick out of the response,” he wrote in 1977.

“It surprised me a bit that Lowry is still number one,” says Wouters. That methods paper has become a kind of symbolic citation, he says: a standardized reference that, for poorly understood cultural reasons, became a norm to acknowledge in particular fields.

But if the Microsoft work or the top-rated DFT paper continue their current vertiginous increase in citations, then they could overtake the Lowry paper before 2030, according to Nature’s analysis of trends from the OpenAlex database.