AI digital twin · reaction-centric design
Predict how a prodrug activates
before you commit every experiment
SPADE Twin is a research digital twin for reaction-centric drug design. It connects molecular structure, activation chemistry, and biological relevance so teams can reason about when, where, and how a candidate may transform - with mechanism in the loop, not only a black-box score.
Early stage · research use only · not a medical device
The problem
Prodrug value is more than parent structure
Designing prodrugs - compounds given in an inactive form and activated in the body - is slow when the facts live in separate places and the models do not talk to each other.
Fragmented knowledge
Identity, mechanism, and related public facts are scattered across different resources and formats. Teams spend time reconciling sources instead of designing.
Disconnected models
Chemistry, reaction kinetics, and biological disposition are often modeled in separate tools. Nothing links structure → activation reaction → downstream relevance in one place.
Trial-and-error activation testing
Because the activation path is hard to anticipate, programs lean on many wet-lab cycles to learn when, where, and how a candidate activates - months of cost for questions a twin could help prioritize.
Our approach
One twin. Three coupled levels.
SPADE Twin is built so classical reaction engineering stays visible inside an AI-assisted framework. The goal is not to replace mechanism - it is to extend predictive reach while keeping physics in the loop.
Molecular level
Represent candidates with structure-based encodings and chemically informed descriptors. Learn non-linear structure-to-reactivity relationships. Surface activation tendency, likely intermediates, and transformation pathways.
Reaction level
Combine learned molecular representations with reaction-engineering principles. Reason about kinetic behavior, rate-limiting steps, and conversion over time - the when, where, and how of activation.
Biological relevance
Connect chemical transformation to downstream processes such as transport, release kinetics, and functional activity - including how molecular choices may shift activation timing and therapeutic relevance for prodrugs.
The result is a rapid virtual-screening posture for transformable therapeutic systems - so experimental work can confirm stronger hypotheses instead of discovering every path from scratch.
How it works
From structure to activation insight
Three editorial steps. No product logins on this site - this is the scientific spine of the twin.
Molecular
- What
- Encode the candidate and predict activation-related chemical behavior.
- Why
- Start from what the molecule is and how it is likely to transform.
Reaction
- What
- Map that behavior onto kinetic and rate-limiting reasoning under reaction-engineering principles.
- Why
- Activation is a process in time, not only a label on a structure.
Biological relevance
- What
- Link transformation to transport, release, and functional relevance for prodrug design questions.
- Why
- Chemistry only helps if it informs what happens in a biological context.
Who it's for
Built for people who design activation
Medicinal / reaction-engineering researcher
- Wants
- Rank prodrug analogs on activation behavior before synthesis.
- Frustration today
- Data and models do not connect chemistry to activation timing.
- Success (vision)
- Choose which analog to make first with a clearer activation hypothesis.
Computational / cheminformatics scientist
- Wants
- Screen candidates and export structured results for the chemistry team.
- Frustration today
- Manual stitching of tools; hard-to-trust black boxes.
- Success (vision)
- Hand chemists a shortlist with transparent multi-level reasoning.
Student, trainee, or PI (light user)
- Wants
- Understand real prodrug activation stories and see a clear summary.
- Frustration today
- Dense tools with no plain-language readout.
- Success (vision)
- Learn or review a candidate path without fighting six disconnected apps.
Where SPADE Twin sits
Structure predictors and docking tools answer different questions. Generic property estimators do not model the activation reaction. SPADE Twin focuses on the transformation from inactive prodrug to active species - and on the timing of that change.
Team
Researchers at Nebraska
SPADE Twin is led by researchers in the Department of Chemical & Biomolecular Engineering at the University of Nebraska–Lincoln.
Request early access
We are building carefully. Join the interest list for research early access updates - no spam, no clinical claims.