Skip to content
HUMAN BIOLOGICAL
SANDBOX

Engine Version One

Open the cell experiment laboratory ↗

Build a digital starting state, create property-size cases, and explore their evolution over a five-year horizon. This page describes the proposed engine architecture.

1. Collect · Build the digital body

Start with the person’s measurements at one test date. Link body parts to tissues, cell populations, and their properties. Each object has an identifier, units, sources, and its own known or unknown values.

Person

Age, weight, height, and test date.

Tissue

Location, measured volume, and applicable cell types.

Cell population

Supported counts, size ranges, and starting properties.

Count a supported cell population

N = V × C, where V is the measured sample volume and C is its sourced cell concentration. For example, 2 mL × 1,000 cells/mL = 2,000 cells. These numbers illustrate the arithmetic, not a normal tissue concentration.

Volume alone does not give cell sizes or DNA states. Missing properties require applicable measurements or reference models.

Count a cell’s building blocks

N = c × V × NA, where c is the entered concentration of a compound, V is the entered volume of one cell, and NA is the Avogadro constant, 6.02214076 × 10²³ per mole. Mass in the cell is m = c × V × M, with M the compound’s molar mass from ChEBI.

For example, 1 mmol/L in a 1,000 fL cell is 0.001 × 10⁻¹² × 6.02214076 × 10²³ ≈ 6.02 × 10⁸ molecules. These numbers illustrate the arithmetic, not a typical cell.

Water and ions

Water and seven inorganic ions, including sodium, potassium, chloride and calcium.

Amino acids and nucleotides

The twenty amino acids of proteins and nine nucleotides of RNA, DNA and energy transfer.

Sugars and lipid components

Glucose, ribose and deoxyribose, with fatty acids, glycerol, choline, cholesterol and sphingosine.

Only these 48 compounds are counted. Any other selected chemical is listed as excluded. Without both a concentration and a cell volume, the engine returns no count. The count describes one entered moment; it does not model how a cell makes, takes up or uses these compounds.

Read the update note · Compound list and molar masses

Explore cell production and existing calculators

2. Run properties · Create cases

Copy the starting database into independent cases. Each case selects plausible sizes or values from a supported range, while preserving the original baseline.

Generate a range of values

For a simple bounded grid, xᵢ = xₘᵢₙ + i × (xₘₐₓ − xₘᵢₙ) / (k − 1), with k ≥ 2 and i from 0 to k − 1. This samples a supplied range; it does not establish which values are biologically possible or equally likely.

Constraints and relationships prevent incompatible combinations. Sampling more cases improves coverage of the chosen model, not proof of complete coverage of a human body.

What “aging” means here

In this project workflow, “aging” means property-size evolution. Chronological person age, property size, and elapsed simulation time remain separate fields.

3. Age · Follow sizes for five years

Version One’s proposed horizon is five years from the starting date. Every scenario runs only the equations applicable to its selected properties and inputs.

Update the state

A model defines dx/dt = f(x, inputs, t). A suitable numerical solver advances its state and checks error and valid bounds. The five-year horizon does not determine the solver’s step size or justify extending a short-term cellular model for five years.

Where no supported equation exists, the value stays unknown or unchanged according to an explicit assumption. The engine must not invent a growth rate.

Keep a reviewable case

Save the baseline link, case identifier, engine release, inputs, sources, equations, trajectory, and defined flagged events. A red flag identifies the model event chosen for review.

Release and validation

The full initializer and connected scenario runner are proposed. Existing server modules calculate cell counts, cell building-block counts and selected research models independently. Version One should become a fixed release after its rules, scope, and tests are agreed; later changes become new releases.

Validate each selected process against independent data before expanding the engine. Different tissues require different models.