Prodius

Four places a better tissue recipe matters.

Drug testing leads, because regulators are asking for human-relevant models now. The workflows below are examples of how the engine would be used. They are not completed projects.

Drug testing

Lead use case

Reproducible human tissue and organoid models for toxicity and efficacy testing, aligned with FDA's shift toward New Approach Methodologies (NAMs). [18, 17, 15]

For

Pharma and biotech safety, DMPK and discovery teams

Example workflow

A human liver tissue model that keeps drug-metabolizing function stable over a multi-week dosing study, batch after batch.

  1. 1Set the acceptance criteria: marker expression, function over time, and allowed batch-to-batch variance.
  2. 2The engine proposes ranked recipes that vary cell source, matrix stiffness, co-culture ratio and media schedule.
  3. 3Replicate batches run on automated hardware and are scored on the same readouts every time.
  4. 4Variance between batches is treated as an outcome to minimize, not noise to average away.

Regenerative medicine

Faster protocol development for cartilage, skin, bone and vessels, and eventually organ tissue. [11, 14]

For

Tissue engineering labs and regenerative medicine companies

Example workflow

Cartilage with hyaline-like matrix and a target compressive stiffness, grown from a defined stem cell source.

  1. 1Define the matrix composition and mechanical targets the implant must reach.
  2. 2Recipes vary hydrogel chemistry, pore size, dynamic compression and the growth-factor timeline.
  3. 3Constructs are measured mechanically and by imaging and sequencing at fixed time points.
  4. 4Each round narrows the protocol window instead of restarting the search.

Disease modeling

Grow diseased tissue, such as fibrotic tissue or a tumor microenvironment, to study mechanisms in human cells. [8, 13]

For

Academic and industry disease biology groups

Example workflow

A fibrosis model whose cell states and spatial organization resemble patient tissue more closely than standard 2D culture.

  1. 1Use reference single-cell and spatial data from patient tissue as the target state.
  2. 2Recipes vary cell mix, matrix stiffness and profibrotic signaling over time.
  3. 3Single-cell and spatial sequencing compare the grown tissue with the reference.
  4. 4The model learns which levers move the tissue toward the disease state.

Personalized medicine

Patient-derived tissue, grown reliably enough to test which treatment works for that patient. [12, 15]

For

Translational and clinical research teams (longer-term)

Example workflow

Patient-derived organoids that grow reliably from a small sample, in time to inform a treatment decision.

  1. 1Start from what the engine has learned across many prior samples of the same tissue.
  2. 2Adapt the recipe to the individual sample with as few runs as possible.
  3. 3Expose the grown tissue to a panel of candidate treatments.
  4. 4Report responses with the uncertainty attached, for clinicians to weigh.

Have a tissue you cannot grow reliably?

Tell us which one and what it has to do. That is how we choose where the engine starts.

Partner with us