The scope of the analysis is unusually broad. The authors surveyed molecular and cellular evidence from 101 multicellular species whose maximum lifespans reach at least 250 years, drawing on the AnAge database and a wide range of primary studies to compile lifespan records. Ninety of these species are plants, from oaks and pines to the ancient seagrass Posidonia oceanica and the creosote bush of the American desert, while eleven are animals, including hydras, sponges, tubeworms and the freshwater planarian Schmidtea mediterranea. Some of these organisms only modestly exceed the 250-year threshold, but others may persist for thousands or even tens of thousands of years. At the far edge of the spectrum, the planarian and the jellyfish Turritopsis dohrnii are considered potentially immortal, meaning that no intrinsic limit to their lifespan has been demonstrated under suitable conditions.

What makes the review methodologically interesting is the way it frames the comparison. Rather than asking only why some species live longer than others, the authors examined two distinct dimensions of longevity: variation between exceptionally long-lived species and shorter-lived ones, and variation among individuals within a single long-lived species. This distinction matters because the mechanisms that make one species outlast another across evolutionary time need not be the same mechanisms that determine why one 500-year-old pine survives its neighbor. The evidence base they assembled spans genetics, genomics, transcriptomics, proteomics and metabolomics, alongside studies of telomere maintenance, stress resistance, regeneration and tumor formation, giving a genuinely multi-layered picture of what extreme longevity looks like at the level of cells and molecules.

Several recurring themes emerged from this cross-species survey. Compared with shorter-lived relatives, extremely long-lived organisms tend to show better DNA maintenance and repair, greater resistance to both cellular and environmental stress, pronounced regenerative capacity, better preservation of stem-cell populations, and mechanisms that appear to suppress cancer development. Comparative genomic studies have additionally identified longevity-associated pathways involved in programmed cell death, protein homeostasis, resistance to oxidative stress and responses to environmental damage. None of these features is exotic in itself; each is a familiar component of cellular housekeeping. What is striking is that they recur together, in organisms as different as a conifer and a cnidarian, suggesting that evolution has converged on a similar toolkit whenever lifespan is pushed to extremes.

Regeneration stands out as particularly important in some of the most remarkable cases. Hydra can continuously replace the cells of its body, effectively renewing itself throughout life. Turritopsis dohrnii can reverse from a mature medusa back to an earlier developmental stage, restarting its life cycle and earning its popular name as the immortal jellyfish. Planarians carry a population of adult stem cells called neoblasts, which can generate many different cell types and support the regeneration of entire body parts, an ability that underlies their apparent exemption from aging. Long-lived plants show an analogous pattern in a botanical register: species such as Scots pine and Ginkgo biloba appear to preserve the pluripotency of their cells during aging better than shorter-lived plants do, maintaining the developmental flexibility needed to keep growing and replacing tissues across centuries.

Equally instructive is what the review does not find. The idea that a single familiar aging mechanism can explain exceptional longevity does not survive the evidence. Telomere length, telomerase activity, metabolic rate and antioxidant activity do not consistently account for lifespan differences among individuals within exceptionally long-lived species. Telomere characteristics vary not only between species but even between tissues within the same organism, which complicates any attempt to read lifespan directly from telomere biology. The authors conclude that extreme longevity arises from multiple interacting biological strategies rather than from a universal molecular switch, a conclusion that cautions against the popular habit of reducing aging to any one measurable marker.

Cancer resistance may be another essential component of the picture. Long lifespans combined with extensive cell renewal should, in principle, multiply the opportunities for malignant transformation, yet tumors appear relatively uncommon in several exceptionally long-lived species. Planarians offer a vivid example: they combine extraordinary regenerative output with robust DNA-repair mechanisms and conserved regulators of cell division, apparently keeping proliferation under control even as tissues are continuously rebuilt. The authors suggest that the anticancer adaptations of long-lived organisms deserve systematic investigation, while being careful to note that the available evidence remains incomplete. Large animals such as elephants and whales have famously evolved extra copies of tumor-suppressor genes, and the new review extends that logic to the plant and invertebrate champions of longevity, where the relevant mechanisms are far less well characterized.

Perhaps the most conceptually significant finding is the convergence between the two dimensions of longevity the authors set out to compare. Mechanisms associated with lifespan variation within species overlap with those that distinguish longer-lived species from shorter-lived ones, and both categories point toward DNA maintenance, stemness and stress-resistance pathways. Notably, these are also the hallmark features of early developmental stages and of germ cells, the two contexts in which biology most conspicuously refuses to age. The authors compress this pattern into a memorable formulation: living young appears to be a strategy of living long. In other words, the longest-lived organisms seem to maintain, into adulthood and old age, a molecular state that most cells abandon during differentiation, keeping their tissues in something closer to an embryonic or germline configuration.

The review also insists that longevity cannot be understood apart from the environment. Temperature, ecological pressures and other external conditions interact with molecular pathways over evolutionary timescales, and these relationships differ substantially among species. A cold, stable deep-sea environment slows metabolism and damage accumulation for an Antarctic sponge, while a desert shrub survives through drought tolerance and clonal growth, so no single environmental condition can be generalized as longevity-promoting. This ecological framing matters for interpretation as well: laboratory conditions rarely reproduce the circumstances under which extreme lifespans evolved, and measurements taken in cultivation or captivity may not reflect what happens in the wild.

For readers hoping that ancient trees and immortal jellyfish will yield a prescription for human life extension, the authors are appropriately restrained. The review does not establish that the longevity mechanisms of these species can be directly transferred to humans. What these organisms provide instead is a set of comparative biological models for identifying pathways that warrant further mechanistic and translational investigation, particularly those linking genome maintenance, stem-cell preservation and stress resistance to healthy aging. Major knowledge gaps remain, and the authors are candid about them: many exceptionally long-lived species have received little molecular study, existing investigations use different methods and tissues, and the field lacks comprehensive cross-species healthspan data comparable to existing lifespan databases. Larger studies using standardized methods and incorporating environmental information will be needed to separate causal longevity mechanisms from mere correlations. Even so, the central message is clear and quietly provocative. Living for centuries does not depend on a single gene or pathway, but on overlapping strategies for keeping DNA intact, preserving regenerative capacity and stemness, resisting stress and limiting cancer. Understanding how those strategies operate across the tree of life may not make humans immortal, but it could redefine what healthy aging means for a species that currently measures its ambitions in decades rather than millennia.

Subject of Research: Comparative biology of exceptional longevity across long-lived plant and animal species

Article Title: Exceptional longevity across species linked to DNA maintenance, stemness and stress resistance

Article References: Exceptional longevity across species linked to DNA maintenance, stemness and stress resistance. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: longevity, aging, DNA repair, stem cells, stress resistance, regeneration, telomeres, cancer resistance, comparative genomics, planarians, Hydra, healthspan

The ROCKET family has since grown rapidly. MiniRocket stripped the original method down to a nearly deterministic transform with a fixed set of 84 kernel patterns containing only the values -1 and 2, dramatically accelerating convolution while keeping accuracy essentially unchanged. MultiRocket enriched the recipe by adding new pooling operators and a first-order difference representation of the input. HYDRA blended random convolutions with dictionary-based ideas, grouping kernels and building histograms of their strongest responses. Across benchmarks, these methods consistently sit at or near the state of the art, and because the only learned parameters are the weights of a linear classifier, they remain among the fastest options available. Yet a closer look reveals a hidden rigidity: nearly all of them commit, at design time, to a fixed input representation and a small, fixed menu of pooling operators.

A new study published in Data Mining and Knowledge Discovery argues that this rigidity is a genuine weakness. Mouhamadou Mansour Lo and colleagues at the University of Artois in France first ran a systematic experiment on the UCR time series archive, the standard benchmark collection for the field. They took MiniRocket and systematically swapped its pooling operator, testing five candidates: the proportion of positive values (PPV), zero crossing (ZC), the mean of positive values, the mean of the indices of positive values, and the longest stretch of positive values. They also varied the input representation, comparing the raw series, its first-order difference, and a concatenation of both. Across 112 datasets and 30 resamples each, the result was striking: no single combination of representation and pooling operator dominated. Even the best average performer, PPV applied to the mixed representation, was beaten on roughly 83 percent of individual datasets by some alternative.

The study did not stop at the aggregate numbers. The authors probed when each pooling operator shines, using synthetic activation maps in which two classes differ in a controlled way. When classes differ mainly in the amplitude of a discriminative pattern, the mean of positive values separates them best. When the pattern’s typical location shifts between classes, the mean of indices of positive values wins. When the pattern’s duration varies by class, the longest stretch of positive values is the strongest discriminator. PPV and zero crossing act as versatile generalists, coping well when a pattern simply appears at different frequencies in each class. A further regularity emerged with series length: PPV tends to excel on short time series, while zero crossing pulls ahead on longer ones, plausibly because longer sequences accumulate richer information in their sign changes. This observation also helps explain why MiniRocket, which relies entirely on PPV, performs so well on the UCR archive, which is dominated by short series.

Motivated by these findings, the team built SelF-Rocket, short for Selected Features Rocket, an extension of MiniRocket that refuses to commit in advance. Instead of fixing the pooling operator and input representation, SelF-Rocket generates features under all fifteen combinations of three input representations and five pooling operators, and then lets the data decide. The heart of the method is a wrapper-based feature selection module. The training set is repeatedly split into stratified train and validation subsets, either through repeated stratified k-fold cross-validation or a stratified shuffle split, depending on dataset size. For each representation-pooling combination, a set of small ridge classifiers, the same linear model used by the original ROCKET, is trained on these splits, and the combination with the highest median validation accuracy is chosen. The median, rather than the mean, guards against lucky runs on unusually easy splits.

One subtlety proved important for small datasets: vote validation. When a dataset contains few examples, the mini-classifiers inside the selection module are trained on very little data, and their votes can be unreliable. SelF-Rocket therefore checks whether the winning combination is broadly supported, requiring it to appear near the top of most voters’ rankings within a configurable threshold. If the consensus is too weak, the method falls back to a sensible default, PPV or zero crossing on the mixed representation, chosen according to the length of the series. Across 112 datasets and 30 resamples, the fallback fired in about 41 percent of runs, and its benefit was most pronounced on datasets with at most 100 training examples, where an ill-validated vote could otherwise steer the classifier badly wrong.

The benchmark results are impressive. Evaluated on the same 112 UCR datasets and the same 30 train-test resamples used in recent community-wide bake-offs, SelF-Rocket improved on its baseline MiniRocket by roughly one percentage point of mean accuracy. Combined with HYDRA’s features, it ranked second only to HIVE-COTE 2.0 among eleven compared classifiers, edging out MultiRocket and Hydra plus MultiRocket on average, and the Wilcoxon signed-rank test found no statistically significant difference between Hydra plus SelF-Rocket, Hydra plus MultiRocket, and HIVE-COTE 2.0. In other words, a method that uses only a single pooling operator at a time matches the accuracy of methods that stack four operators, while enjoying a practical advantage at prediction time: SelF-Rocket feeds its classifier only 9,996 or 19,992 features, depending on the representation, compared with roughly 50,000 for MultiRocket, which translates into faster classification.

The authors are candid about the costs and the remaining headroom. Feature generation is heavier than MiniRocket’s, since roughly 9,996 kernels are instantiated per representation type, and the selection module adds training time, though with the mixed representation and default settings the module takes under two seconds per resample on average. Ablation studies showed that removing any single pooling operator does not significantly change accuracy, but removing PPV hurts more than removing the longest-stretch operator, and the mixed representation generally outperforms either raw or differenced inputs alone. To quantify the ceiling, the team also ran an oracle version of SelF-Rocket, in which the best representation-pooling pair is known in advance for each dataset. The oracle’s results sketch how much performance a better selection module might one day unlock, a question the authors leave explicitly open.

The implications reach beyond one benchmark. Random convolution kernel methods are already prized in domains from healthcare to industrial monitoring precisely because they are fast, simple, and accurate; making them adaptive removes a design decision that practitioners would otherwise have to guess. The implementation is open source on GitHub, and the authors outline clear next steps: richer input representations such as second-order differences, Fourier or Hilbert transforms, faster filter-based selection using chi-squared tests, and improved vote validation to close the gap to the oracle. For a field that has spent years squeezing accuracy out of fixed recipes, the message of this study is refreshingly simple: the best features depend on the data, and a classifier that checks, rather than assumes, is measurably better for it.

Subject of Research: Adaptive feature selection with random convolution kernels for time series classification

Article Title: Time series classification with random convolution kernels: pooling operators and input representations matter

Article References: Lo, M. M., Morvan, G., Rossi, M., Morganti, F., & Mercier, D. (2026). Time series classification with random convolution kernels: pooling operators and input representations matter. Data Mining and Knowledge Discovery, 40(6), Article 100. https://doi.org/10.1007/s10618-026-01269-w

Image Credits: AI Generated

DOI: 10.1007/s10618-026-01269-w

Keywords: time series classification, random convolution kernels, ROCKET, MiniRocket, pooling operators, feature selection, machine learning, SelF-Rocket, UCR archive, ridge classifier, HYDRA, MultiRocket