Digital Twins for Longevity: Dr. Dmitrii Chebanov on Predicting Your Health
Computational biologist Dr. Dmitrii Chebanov explains how a 'digital twin' — a living computer model of your body built from DNA, biomarkers and daily symptoms — can simulate how interventions will affect you months or even years ahead. The aim is to shift health from reactive to proactive, letting you test supplements and habits virtually before committing months of your life to them.
Overview
Dr. Dmitrii Chebanov, a computational biologist at Memorial Sloan Kettering Cancer Center and chief scientist at Holivita, joins Modern Healthspan to explain how computer models of the human body could make longevity practical. He describes a 'digital twin' as a computational representation of a person, built from their biological and clinical data, that can simulate how the body reacts to nutrition, activity and supplements before anything is tried in real life.
The approach grew out of cancer drug development, where digital twins of patient cohorts helped design smarter clinical trials; Holivita adapts the idea to the far more complex problem of whole-body longevity in a single individual. Chebanov walks through the platform's three data layers — DNA as the body's static 'architecture', dynamic biomarkers including epigenetic methylation that show how that architecture performs, and subjective symptoms like sleep, energy and mood. He makes the case for whole-genome sequencing over cheap panels, noting that new genetic associations are discovered constantly and a full genome lets you revisit your data as science advances.
Rather than showing users raw numbers, the platform aims to explain what is changing and what to do about it, pairing a personal dashboard with a three-month strategy that is recalculated over time. He also highlights data ownership and portability — users own their health data, can carry it to any clinician, and may choose to share it with researchers. The long-term vision is a continuously updated model of each person's health that moves care from reactive to proactive and predictive. Throughout, he frames the tools as support for everyday decisions, never a replacement for medical care.
Key quotes
5When you know the mechanisms a supplement impacts, we can predict how they will behave in three months, or even in ten years.
Instead of looking only at the physical patient, we create a computational representation of that patient — a computer model of a person.
DNA is the architecture of our body. It says where the strengths are and also where the vulnerabilities are.
The first human genome sequencing cost about five billion dollars and took thirteen years; now we can do it in six weeks for, let's say, two hundred dollars.
Everyone should have a continuously updated model of their own health. You can call it a digital twin.
Key ideas
9A companion that prioritizes what actually works
With an overwhelming number of available interventions, the platform is designed to help each person separate what truly works from marketing and myth, and to give them real control over their health.
What a digital twin is
A computational representation of a person, built from biological and clinical data, that can simulate the body's reactions to habits, nutrition, activity and supplements before anything is tried in real life.
Lessons from cancer trials
Because more than 90% of clinical trials fail, digital twins of patient cohorts were used to help clarify the right dose, drug, patient and timing — reducing costly failures before a trial even began.
From cohorts to individuals
Cancer modeling averages across a cohort, while longevity requires modeling one specific person — a far more complex system spanning predictions decades into the future.
Simulating interventions instead of trial-and-error
Rather than testing one supplement at a time over months, the platform simulates how roughly a hundred actionable supplements interact so combined outcomes can be projected ahead.
Five biological systems
Aging is decomposed into systems such as immune function, inflammation ('inflammaging') and protein interactions, modeled separately so customers do not need to be scientists themselves.
Three layers of data
DNA is the static architecture, dynamic biomarkers show how that architecture actually performs, and subjective symptoms capture how a person feels day to day.
Why whole-genome sequencing
New gene variants and associations are discovered all the time; a full genome lets you keep returning to your own data as science advances, without sequencing yourself again.
Measure, understand, act, measure again
You connect the data you already have, see a predicted trajectory that helps you stay motivated, and the system recalculates your strategy roughly every three months.
Practical takeaways
6- 1
Rethink one-at-a-time testing 22:30
Instead of waiting months to judge a single supplement, mechanism-based simulation can model combinations and project their effects ahead of time.
- 2
Favor whole-genome sequencing 34:00
If you sequence your whole genome once, you can keep mining it as new discoveries appear, rather than re-testing for each new marker.
- 3
Track dynamic biomarkers, not just DNA 30:30
A genetic predisposition only matters if your biomarkers show it is actually realized, so measuring over time is what makes the data actionable.
- 4
Log how you actually feel 44:00
Sleep, energy, digestion, mood and focus are treated as a core data layer — subjective markers, not an afterthought.
- 5
Keep your health data portable 56:00
Storing comprehensive health data in one place you own lets any clinician pick up where the last one left off, avoiding costly re-diagnosis.
- 6
Move from reactive to proactive 1:02:00
The goal is to understand what is happening in your health earlier and act on early signals before small changes become problems.
Topics & chapters
15Cold open: predicting health years ahead
Chebanov previews mechanism-based prediction and the digital-twin vision before the interview begins.
What Holivita is
A companion that helps each person control and enhance physical and mental health amid a flood of possible interventions.
What is a digital twin
A computer model of a person, built from biological and clinical data, that simulates how the body reacts.
Digital twins in cancer trials
How cohort models helped design clinical trials and cut the high failure rate during his work at Memorial Sloan Kettering.
Cohorts vs individuals
Moving from averaged cohorts to a single, far more complex person over a 20–30 year horizon.
Where is the outcome data
Chebanov says individual-level results are being compiled and he hopes to present them to the scientific community.
Testing many interventions at once
With roughly a hundred actionable supplements, simulation replaces slow, one-at-a-time trials.
Five biological systems
Immune function, inflammation, protein interactions and more, modeled so customers need not be scientists.
Three layers of data
DNA, dynamic biomarkers and subjective symptoms together build the model of your health.
Why whole-genome sequencing
Full sequencing lets you revisit your data as new variants are found, and costs keep falling fast.
Epigenetics, proteome and microbiome
Methylation helps estimate biological age; proteomics and microbiome await standardized methods.
mRNA and transcriptomics
Gene expression shows which genes are active right now and offers a way to act on mechanisms.
The user experience
Connect everything you have, then measure, understand, act and measure again, with a dashboard and chatbot.
Setting a goal and a portable record
From 'just get healthier' to a three-month strategy, plus a health record any clinician can use.
Data ownership and the future
You own and can choose to share your data; the vision is proactive, predictive health.