Buck Institute perspective argues for longitudinal, AI-driven detection of disease risk years before symptoms appear.
Most chronic diseases arrive late to clinic but early to biology. Long before a diagnosis of type 2 diabetes, cardiovascular disease or neurodegeneration is written into a medical record, molecular drift has often been underway for years – sometimes decades. A new perspective from researchers at the Buck Institute for Research on Aging argues that medicine has been listening for the shout, when it should have been attuning itself to the whisper [1].
Published in Cell, the paper proposes a conceptual and computational shift: away from single-cause narratives and population averages, toward longitudinal, multimodal monitoring interpreted against an individual’s own baseline. The authors call this distributed, cumulative influence the “long tail” of biology – a terrain of small, context-dependent perturbations that, taken together, shape aging trajectories and disease risk [1].
The long tail of risk
The paper takes aim at a familiar story in biomedical research – that disease can be traced to a dominant driver. BRCA1, APOE, LDL receptor mutations; important, certainly, but not representative of how most chronic illness unfolds. As the authors write, “Most disease risk, and especially aging itself, does not stem from a single cause but from a complex interplay of subtle, cumulative changes across physiology [1]”.
Type 2 diabetes serves as a case study. Hyperglycemia is the diagnostic threshold, but inflammatory shifts, changes in fat distribution, microbiome alterations and declining beta-cell resilience can precede that threshold by 10 to 15 years. An early warning system with bells on. While blood glucose is the clarion call, the earlier changes are quieter, more diffuse and harder to interpret. Neurodegeneration and cardiovascular disease follow similar arcs – early molecular or structural deviations, then clinical manifestation.
“By the time many diseases are diagnosed, the body has often been drifting off course for years,” said Nathan Price, PhD, senior author of the paper. “We now have the opportunity to detect those early changes by tracking what’s normal for each individual and noticing when biology starts to move in the wrong direction.”
The proposed solution is what the authors term an “N-of-1 analyzer” – a framework that treats each person as their own biological reference, using longitudinal multi-omics, wearable data and AI-based reasoning to identify divergence from personal baselines. It is not merely anomaly detection; the ambition is mechanistic interpretation, linking shifts in analytes, networks and trajectories to plausible causal pathways.
Longevity.Technology: The Buck team’s “biology whispers” framing is more than a clever turn of phrase; it is a very geroscience way of saying that chronic disease is often the end-stage paperwork of a much longer process in which resilience erodes quietly, system by system, year by year. What’s interesting here is not another plea for “more biomarkers”, but a philosophical shift away from single-cause hero narratives toward the long tail of small, context-dependent perturbations – and the uncomfortable implication that prevention will look less like a one-off test and more like longitudinal surveillance, interpreted against you rather than the population average. Done well, this could sit alongside today’s aging toolkits: epigenetic clocks as relatively robust, slower-moving risk stratifiers, multi-omic and wearable signals as the twitchier but potentially more actionable early-warning layer. Done badly, it becomes an expensive hobby that generates exquisite data, anxious patients and clinician burnout. The authors are candid about that translational gap – cost, noise, reimbursement and the regulatory gymnastics of AI-driven interpretation – and that honesty matters, because the technology is arriving faster than the health system’s ability to decide who gets to listen, how often and what, precisely, we do when the whisper turns out to be a false alarm.
From clocks to trajectories
The paper situates this proposal within the broader landscape of aging biomarkers [1]. Epigenetic clocks, now widely studied, offer relatively robust associations with morbidity and mortality risk; they are, however, often slower to respond to short-term intervention and can lack mechanistic clarity. Multi-omic measures of biological age and related constructs such as “biological BMI” aim to capture more dynamic shifts in inflammation, metabolism and immune function, potentially offering a more actionable readout.
Both approaches have limitations. Tissue specificity, population generalizability and cost remain open questions. Yet the authors argue that the integration of these measures into longitudinal profiles – rather than their use as static scores – is where preventive potential may lie. Aging, in this view, is not a number but a trajectory.
Explaining that medicine has traditionally focused on treating disease after symptoms appear, lead author Noa Rappaport, PhD, said: “Our goal is to shift toward protecting health by identifying risk earlier and understanding how each person’s biology changes over time.”
Technological feasibility is advancing quickly. High-dimensional plasma proteomics, metabolomics and cell-free DNA assays can now profile thousands of analytes from small samples. Non-invasive biosamples – like urine, saliva, hair and breath – offer complementary windows into metabolism and exposure, and wearables provide continuous physiological data streams, capable of flagging deviations from personal norms before symptoms arise. Data are no longer the bottleneck.
Interpretation is harder. The authors devote substantial attention to signal-to-noise management in high-dimensional monitoring – false positives, alert fatigue and the compounding risk of spurious findings when thousands of variables are tracked longitudinally. Tiered reporting, temporal validation and biological coherence filters are proposed as guardrails [1].
Economics and equity
While the vision is rightly ambitious, the economics are less forgiving. Comprehensive multi-omic profiling can cost thousands of dollars per assessment, and insurance models remain structured around acute care rather than longitudinal prevention. Without careful implementation, the authors acknowledge, personalized prevention could become a premium service.
“Advanced biological testing can still be expensive, and healthcare systems are largely designed to treat illness rather than monitor long-term health,” said Lee Hood, MD, PhD, distinguished professor and co-director of the Buck’s Center for Healthspan.
“Ensuring broad access to preventive technologies will be critical to preventing new health disparities. In addition, regulatory systems will need to adapt to evaluate new approaches that rely on personalized data and AI-driven analysis.”
The regulatory challenge is equally complex. Adaptive AI systems and digital twins do not fit neatly into traditional drug or device pathways. The authors suggest a phased approach, beginning with high-complexity laboratory-developed tests before progressing toward more adaptive, learning systems as evidence accumulates.
This is not a call for mass screening tomorrow. It is a proposal for architectural redesign – of data collection, clinical workflow, reimbursement logic and regulatory science.
A quieter clinic
In an aging world, the burden of chronic disease is less a sudden storm than a slow weather system. But if medicine is to move upstream, it is going to have to tolerate ambiguity, manage complexity and invest in resilience before pathology becomes irreversible. Listening earlier will not be simple and it may not be cheap, but it could, over time, change what we mean by prevention – and what we expect from the clinic.
Photographs courtesy of The Buck Institute for Research on Aging
[1] https://www.cell.com/cell-systems/abstract/S2405-4712(25)00342-4
