To help identify and prioritise the more pressing open problems in longevity and ageing science we combined community input, data-driven analysis and manual review (see Methods). All open problems were collected through an online platform as well as a dedicated workshop which ensured engagement from the scientific community. We assessed the open problems by applying NLP techniques to analyse their prevalence in the PubMed database. This allowed us to quantify the extent to which an open problem may have been explored, highlight potential gaps in research and inform us on the final selection of open problems. Alongside this, we utilised clustering methods to help us group our open problems into appropriate categories for easier exploration for users on our website.

Data collection of open problems

The number of open problems we collected through submissions via the website and through the workshop totalled to 290. Online submissions totalled to 160 open problems while the workshop produced 130. Initial pre-filtering of questions before grouping and NLP analysis reduced this to a total of 204. These 204 open problems were then used for clustering and NLP analysis against PubMed articles.

NLP analysis of open problems using PubMed literature

Collection of all PubMed articles under the MeSH term of “Ageing” totalled to 389,627 articles. After the removal of articles that did not contain an abstract, the number of unique articles totalled to 200,228 (a 48.6% reduction) articles which were then paired to the 204 selected open problems. This created a total of 40,846,512 pairs of open problems and articles to be analysed. Using the PubMedBert embeddings of the open problems and article texts, we assessed the relationship between the open problems and the PubMed articles on ageing.

The NLP analysis of the 204 open problems revealed variability in their representation within the ageing research literature (Supplementary Table 1). After the application of the language models, a total of 172,031 relevant article and open problem pairs were identified. The representation of individual open problems ranged from 1 to 10,808 articles, with a median of 437 and a mean of 847 (95% confidence interval, CI, calculated at 685–1010) articles per problem. The standard deviation of 1176.7 reflects substantial disparities in research focus across different open problems in ageing research.

Summary of the top and bottom open problems

A statistical analysis of the top and bottom 20 open problems ranked by the total number of associated articles shows a disparity in their representation within the ageing research literature (Supplementary Table 1). The top 20 open problems collectively account for 69,322 articles, representing 40.3% of the total dataset, with an average of 3466.1 (95% CI 2552–4381) articles per problem. In contrast, the bottom 20 problems account for only 341 articles, 0.2% of the dataset, with an average of just 17.05 (95% CI 11.7–22.4) articles per problem.

The top 20 problems dominate the literature, reflecting well-established questions regarding the causes of ageing and even whether fundamental ageing processes exist (Table 1). These problems align closely with current research priorities and methodologies, such as cellular and molecular mechanisms of ageing, evolutionary and comparative biology, the role of model organisms in ageing research, biomarkers and anti-ageing interventions, reinforcing their centrality to the field. Article counts for these problems range from 2196 to 10,808, showing the breadth of interest in these topics.

Table 1 Examples of open problems with significant representation in the literature, categorised by topic

In contrast, the bottom 20 open problems have article counts ranging from 1 to 36. These less explored problems often focus on emerging or niche areas, including methodological challenges, novel therapeutic approaches and unexplored biological mechanisms (Table 2). Naturally, more underrepresented open problems tend to be less supported by existing literature. Nonetheless, these often more specific topics can provide tangible research goals in a shorter time frame (Table 2).

Table 2 Examples of open problems with the least representation in literature, categorised by topicFinal themes and distribution of open problems

The final list of 100 open problems was grouped into 11 themes, ensuring thematic alignment. This grouping process facilitated the organisation of problems into a logical structure that reflects the breadth of the open problems. Each problem was assigned to one primary theme that best fitted its focus.

Figure 1 presents the distribution of the 100 open problems across the 11 themes. The largest proportions were assigned to broader Ageing Mechanisms, more specific Molecular Mechanisms, and Interventions, which collectively accounted for over half of the selected problems. Themes such as Environmental and Physical Factors and Diversity in Human Ageing were less represented, which may reflect they are less explored topics in the field. The final list of 100 open problems is available as Supplementary Table 2.

Fig. 1

Pie chart of the total number of open problems distributed across the 11 themes

Presentation of the final 100 open problems

To maximise engagement and accessibility, the final list has been published on the Longevity Knowledge App website (www.longevityknowledge.app). The website serves as an interactive platform, enabling researchers and the public to explore the open problems and their themes. Each problem is presented with detailed information, including:

Title of the open problem.

Theme to which it belongs.

Metadata, such as genes or compounds, that link to external databases.

Links to related open problems.

A section where users can post any solutions or proposals that can tackle the open problem.

Our website is designed to stimulate further discussion, collaboration and roadmap planning in the ageing research scientific community.

Discussion

Asking the right questions is crucial to advance a scientific field and steer its future directions. As such, we took inspiration from the pioneering work of Strehler to identify, through a collective and systematic effort, a new set of 100 open problems in ageing science. Our list of open problems purposedly includes a combination of broad, big picture questions and specific questions. While broad questions, for example regarding why we age, set long-term prospects in the field, more specific questions may be answerable in the foreseeable future. In other words, we need both big picture questions—for direction—and precise questions to advance biogerontology, which our work provides. Besides, while some topics are more represented than others, for example questions on mechanisms of ageing are the most popular, we also aimed to have a diverse set of topics represented in our final list of questions. As such, our open problems aim to stimulate further discussions and provide a cornerstone in ageing research for years to come. Moreover, given the dynamic nature of the field, our website can be used in the future to update the list of open problems to reflect evolving priorities.

The topics covered in our open questions reflect those we received on our website and the discussions during our workshop. While we made efforts to avoid systematic biases and be inclusive in our list of questions, for example by having a public website anyone in the world could have contributed to, our list will still reflect biases in how contemporary scientists perceive ageing and those topics they find more important. As such, it is not surprising that a major emphasis is on mechanisms of ageing and understanding why we age. This is still a key but major open question in the field as the drivers of ageing remain open to debate18. Many open questions focus on processes or mechanisms hypothesised to be associated with ageing, again ranging from broad questions to more specific ones that are amenable to experimentation. We hope some of our more specific questions broaden the field’s research directions and are useful for students and researchers to address in the next five years.

Another broad topic of great interest is developing interventions targeting ageing. Again, there is significant discussion concerning longevity therapeutics, regarding both the effectiveness of existing potential therapies and how to test them in a clinical setting. It is clear, however, that there is great interest in developing effective interventions for ageing and this is one of the major areas for demonstrating clinical efficacy in the future5. Several questions also concern biomarkers, e.g. which would be the most suitable for evaluating therapies. Questions also arose regarding the nature of existing biomarkers, such as epigenetic clocks, the genetic and environmental determinants of ageing in humans and model organisms and questions regarding recent methods such as partial reprogramming. A few questions concerned the nature of the ageing process, which others have also debated without reaching a consensus12,13, for example regarding cell autonomous and systemic contributors to ageing across tissues and organs. A subset of questions also concerned specific organs and tissues, such as the immune system, and the relationship between ageing processes and disease.

Evolution and comparative biology were another broad topic of interest, again a historically important topic that has not been resolved. Broadly speaking, we still do not understand species differences in ageing, and therefore, it remains another major open area for research. Despite recent progress in finding common ageing markers across mammalian species19, the level of conservation of ageing mechanisms across species remains under debate.

We did not use the initial set of open questions by Strehler, now nearly 50 years old14, as a basis for our work. Hence, how do our 100 open questions compare to those of Strehler? Broadly speaking, some topics remain the same, such as the role of genetics in ageing and longevity, how transcription and translation relate to ageing, the timing of development in long-lived species, species differences in ageing, and even gene therapies for ageing. A few of the questions from Strehler have been partially addressed, for example regarding the genetics of longevity in animal models6, but most remain open. There is also an overlap in mechanisms of ageing between Strehler’s list and ours, for example regarding mitochondria, cellular changes and the role of the immune system. In terms of differences, Strehler gives much more emphasis to CNS diseases, including not only neurodegenerative conditions but also others such as dyslexia and Batten’s disease, which are absent from our list. Furthermore, our list is shaped by modern priorities, including a strong emphasis on developing and evaluating interventions—such as senolytics, partial reprogramming and establishing biomarkers to guide clinical application. This shift may reflect a broader evolution in the field, from characterising ageing changes to actively seeking to modulate them. Our list also reflects a deeper knowledge of molecular biology. For example, Strehler referred to enzymes, tRNAs and ribosomes—broader aspects of biology that are rarely now studied in the context of ageing, even though their role remains an open question. With the advent of molecular biology and large-scale-omics analyses, we should perhaps take notice of old-fashioned questions for which we still lack answers. As such, we see our list as reflecting advances in the field, but complementary to those of Strehler. We have no doubt that future lists of open questions in biogerontology will only partly overlap with ours. While asking the right questions is essential to advance science, answering—or attempting to answer—them often leads to more questions.

In conclusion, our work aims to provide future directions for ageing research that will allow us to address key challenges that remain unanswered. Our open questions reflect current large discussions and unknowns in the field of biogerontology. In spite of progress, for example in developing interventions that retard ageing in preclinical models, there is still huge debate regarding the underpinning mechanisms of ageing, the nature of biomarkers of ageing and a route to developing effective interventions for ageing in humans. Our questions reflect current key disagreements in the field (Table 1), and we hope will serve as inspiration and guide to researchers as well as lay a path for advancing biogerontology.