Launching "Beyond Life": Building Our First Digital Cell

Aging is often described in metaphors: a clock, a slow fire, a gradual loss of order. But if we want to intervene in aging in a systematic way, metaphors aren't enough.

At Systems Beyond, we're starting the aging project by doing something very concrete:

We are building a "digital cell" — a computational model of a somatic human cell that we can expose to different conditions and see how it ages, adapts, or breaks.

Why start with a virtual cell?

Aging is a systems problem. It is not just DNA damage, or mitochondrial decline, or chronic inflammation, but the interaction of all those processes over time.

Experimentally, it's hard to isolate and combine all possible factors. You can't systematically expose the same cell to 100 different "lifetimes." You can't fast-forward decades of biology in a lab.

A computational model, even a simplified one, can run "what if?" scenarios in minutes instead of years, combine multiple stressors (nutrient excess, DNA damage, inflammation) in controlled ways, and let us test hypothetical therapies without touching a real cell yet.

We're not trying to replace biology. We're trying to build a sandbox where ideas can be tested, refined, and prioritized before going back to the lab.

What is our "digital cell"?

For this first stage, our digital cell is a proliferative mammalian somatic cell — think of it as a simplified fibroblast/epithelial cell with the main aging-relevant pathways wired in.

It is a logical network of key modules:

• Nutrient and energy sensing — Insulin/IGF, mTOR, AMPK, sirtuins
• Damage response and senescence — DNA damage response, p53, p21, p16/Rb, senescence program
• Stress and proteostasis — ER stress, unfolded protein response, autophagy
• Mitochondria and oxidative stress — mitochondrial function, ROS, Nrf2
• Inflammation and SASP — NF-κB, JAK–STAT, senescence-associated secretory phenotype
• Cell fate decisions — proliferation, quiescence, senescence, apoptosis

We will gradually add more detail, but v1 is intentionally minimal. The goal is clarity first, complexity later.

How we encode the biology: Boolean networks first

To keep the first model transparent and easy to share, we use a Boolean network — an "on/off" logic model. Each node (e.g. mTORC1, Autophagy, Senescence) can be active (1) or inactive (0), and each node updates according to a logical rule based on other nodes and external inputs.

For example, in this simplified logic:

• mTORC1 turns on when nutrients are high and AMPK is low.
• Autophagy turns on when AMPK or sirtuins or FOXO are active, and mTORC1 is low.
• Senescence turns on when p53 stays high under chronic damage or inflammation.

We define a small set of external inputs that you can "dial in" to run scenarios: nutrients, DNA damage, ROS, ER stress and inflammatory signal. Everything else in the network updates based on these and on each other.

This may look simple, but Boolean networks can already show multiple attractor states (a "healthy cycling" state vs a "senescent + SASP" state), phase transitions where small changes flip the system into a new regime, and how combinations of interventions might push the cell back from a "bad basin" into a more youthful, resilient regime.

Open science: how the Systems Beyond lab will work

This is an open science project. The model structure — nodes, rules, diagrams — will be public. We will share versions over time so you can see how the model evolves, and we will post notebooks, code snippets and visualizations as we go.

You don't need to agree with our choices. In fact, we want people to fork, criticize and improve the model. If you'd like to participate, reach out through the contact form.

This is the beginning of Beyond Life: a shared, evolving model of how a single cell moves through time, stress and repair — and how we might nudge that trajectory toward a longer, healthier life.