chatGPT:
Here’s a concise review of the uploaded paper, Reconstructing mammalian lifespan evolution reveals strong phylogenetic effects and lifespan-associated genes.
Summary
The paper asks two main questions: how mammalian lifespan evolved across the phylogeny, and which genes are associated with those lifespan differences. Using lifespan and body-mass data for 968 mammalian species, the authors normalize lifespan by body size using the longevity quotient (LQ), then test multiple models of trait evolution. They find that lifespan is strongly phylogenetically structured, with Pagel’s λ giving the best fit and λ estimated at 0.96, implying that shared ancestry explains much of the variation in relative lifespan across mammals. They reconstruct ancestral states and conclude that relatively short lifespans evolved repeatedly across many mammalian lineages, whereas relatively long lifespans are more clustered, especially in primates and bats.
They then move to comparative genomics. Across 122 mammals with genomic and lifespan data, they analyze 15,231 one-to-one orthologs and identify 628 genes whose relative evolutionary rates correlate with relative lifespan: 396 negatively correlated genes and 232 positively correlated genes. Negatively correlated genes are enriched for DNA repair, DNA damage response, NF-κB-related functions, and cell division/cell migration, suggesting that stronger evolutionary constraint on these pathways is associated with long-lived mammals. Positively correlated genes are enriched for muscle-related processes and calcium ion transport, which the authors interpret as genes more likely to show relaxed constraint or adaptive change in long-lived species.
They next examine selective regimes. In long-lived mammals, positively correlated genes are much more likely to show relaxed selection, while negatively correlated genes are relatively more constrained and less often positively selected. This leads the authors to argue that mammalian longevity has been shaped by both intensified constraint in some pathways and relaxation in others, rather than by a single evolutionary mechanism.
Finally, they perform a functional screen in C. elegans on 11 candidate genes with orthologs in the worm. Most had no lifespan effect, one candidate shortened lifespan when knocked down, and knockdown of phi-53 extended lifespan. That intervention also improved age-related worm phenotypes, upregulated immune-related genes, downregulated reproduction-related genes, and greatly impaired egg laying, which the authors interpret as consistent with a reproduction–soma trade-off.
What seems novel
The strongest novelty is scale. The study claims the largest taxonomic sampling yet for this question, using 968 mammalian species for the macroevolutionary analysis and substantially expanding genomic sampling relative to an earlier 61-species comparative study by adding 81 unique species across additional genera, families, and orders. That matters because lifespan is phylogenetically non-independent, and broader sampling should improve stability of comparative inference.
A second novelty is conceptual. Rather than looking only for genes under stronger purifying selection in long-lived mammals, the authors explicitly analyze both negatively and positively correlated genes and show that relaxed selection is also a major part of the signal, especially for the positively correlated set. That is a useful shift away from the simplistic idea that longevity genes should all look more conserved in long-lived species.
A third novelty is the phylogenetic framing. They do not just correlate genes with lifespan; they first test alternative models of trait evolution and show that Pagel’s λ fits best. That strengthens the paper methodologically because it foregrounds phylogenetic structure rather than treating species as quasi-independent datapoints.
A fourth novelty is the link to a poorly characterized gene, phi-53, in the worm validation arm. The paper connects inhibition of this gene to lifespan extension plus a transcriptomic signature suggestive of increased immune/stress response and reduced reproductive investment. That is potentially interesting as a hypothesis-generating lead.
Critique
The paper is ambitious and useful, but I think its evidential strength is uneven.
First, the macroevolutionary part is stronger than the mechanistic part. The phylogenetic signal result is convincing: lifespan clearly is not independent across mammals, and the Pagel’s λ fit looks robust. But the gene-level conclusions are still correlational. Relative evolutionary rate associations do not show causation, and enrichment results can be driven by broad pathway properties rather than direct roles in aging. So the mammalian comparative part is best read as a prioritization framework, not a direct map of longevity mechanisms.
Second, the genomic analysis uses only 122 species, far fewer than the 968 used for the trait analysis. That is understandable because of genome availability, but it means the strongest phylogenetic claims and the gene-association claims are being made on different datasets. The authors argue the 122-species sample is representative, but this still leaves room for sampling distortion, especially if unusual clades are over- or underrepresented in the genomic subset.
Third, some thresholds are arbitrary. The authors define exceptionally short- and long-lived lineages using the bottom and top 5% of LQ values, corresponding to LQ 0.44 and 2.40, and then use these cutoffs in downstream selection analyses. They acknowledge the cutoff choice, but this kind of dichotomization can alter which genes appear under relaxed or intensified selection. A sensitivity analysis across several thresholds would have strengthened confidence.
Fourth, the worm validation is only partial. Out of 11 tested genes, most had no detectable lifespan effect, one shortened lifespan, and one extended it. That does not invalidate the mammalian screen, but it does suggest that many comparative hits may not translate cleanly into worm lifespan phenotypes, either because of biology or because the orthology/function mapping is weak. Also, RNAi knockdown in worms is a blunt intervention compared with subtle evolutionary tuning across mammals.
Fifth, the phi-53 result may be biologically interesting but is not yet a strong mammalian longevity mechanism. The lifespan extension is accompanied by sharply reduced egg laying and downregulation of reproductive programs, so a classic fecundity-longevity trade-off is a very plausible explanation. That does not make the result unimportant, but it means the paper has not shown that phi-53 improves lifespan by a mammal-relevant anti-aging mechanism rather than by pushing worms into a low-reproduction state.
Sixth, the manuscript is still “article in press” and explicitly says it is an unedited early version. So some wording, statistics, and even minor inconsistencies may yet change before final publication. One small example is that different parts of the text refer to 13 versus 16 long-lived mammals in the foreground set, which may be a drafting or editing issue but still undermines precision a little.
Seventh, the paper’s own limitations section is fair and important: taxon sampling is still incomplete, some lineages are underrepresented, and the phylogenies come from public databases with topological uncertainty. Those issues matter because ancestral reconstruction, phylogenetic signal estimation, and relative-rate inference all depend on the tree.
Bottom line
This is a strong comparative-evolution paper and a weaker mechanistic-aging paper. Its most persuasive contribution is showing that mammalian lifespan is deeply shaped by phylogeny and that longevity-associated genes include both highly constrained pathways, especially DNA repair and cell-cycle-related functions, and more relaxed pathways, especially calcium transport and muscle-related functions. Its weakest point is the jump from cross-species evolutionary association to mechanistic claims, especially when the experimental validation is mostly in worms and the standout hit also strongly affects reproduction.
If you want, I can next do a deeper “novelty versus prior literature” comparison against the 2020 eLife pan-mammalian lifespan paper.