MAT Transporter
Predict dopamine, norepinephrine, and serotonin transporter activity. The pharmacological fingerprint of every ADHD stimulant.
0.968 AUC
Built for ADHD drug development.
From transporter profiling to PK/PD.
DAT/NET/SERT profiling, abuse liability scoring, BBB permeability, and lisdexamfetamine PK/PD simulation. The only platform purpose-built for stimulant pharmacology.
ML is the solvent.
CC(N)Cc1ccccc1
Try it free
Paste a SMILES string to predict transporter activity, abuse liability, and ADMET properties
The Platform
End-to-end computational pharmacology for ADHD and stimulant drug development
Dis-Solved combines stereo-aware graph neural networks with validated pharmacological rules to profile the targets that define ADHD pharmacology: dopamine and norepinephrine transporter activity, CNS penetration, abuse potential, cardiac safety, and metabolic stability. All from a single SMILES input.
Input
SMILES string
Dis-Solved Engine
GNN + stimulant pharmacology rules
ADHD-Ready Profile
Screen / advance / schedule
- DAT/NET/SERT transporter profiling for stimulant target engagement
- Abuse liability scoring to de-risk scheduling conversations
- BBB permeability with stereoisomer awareness for CNS access
- Lisdexamfetamine PK/PD simulation validated across 67 subjects
- hERG cardiotoxicity and CYP450 metabolism screening
- Full ADMET profile from a single SMILES input
BBB Predictor
Blood-brain barrier permeability prediction with stereochemical encoding
ADHD drug candidates must cross the blood-brain barrier to reach dopaminergic and noradrenergic targets in the prefrontal cortex. Our stereo-aware GNN encodes 3D molecular geometry directly into the prediction, capturing chiral and conformational effects that descriptor-based methods miss. Validated on external holdout data with state-of-the-art performance.
0.96 External AUC
0.92 Internal AUC
Stereo-Aware GNN
Try BBBP Predictor
ADMET Toolkit
ADHD-relevant ADMET profiling from a single SMILES input
Profile the pharmacology that matters for stimulant drug development: monoamine transporter engagement, abuse potential, cardiac safety, and metabolic liabilities.
Abuse Liability
HIGH/MODERATE/LOW scoring based on MAT profile and structural patterns. Built to flag scheduling risk early in the stimulant design cycle.
Multi-classCardiotoxicity
hERG channel blocking prediction. Critical for stimulants, where cardiovascular monitoring is already standard of care.
0.91 AUCCYP450 Metabolism
Drug-drug interaction risk across major CYP isoforms. ADHD patients often take multiple medications; CYP profiling catches co-prescription liabilities.
0.88 AUCADHD PK/PD Simulation
DoseTrack
Mechanistic pharmacokinetic simulation for lisdexamfetamine (Vyvanse)
Log a dose. See the curve.
9.1% Cmax MAPE
Michaelis-Menten + 2-cmt
RK4 ODE Solver
67 subjects validated
- Peak concentration predicted to 9.1% MAPE across 6 independent dose-dataset combinations.
- Food effect derived from first principles. Correctly orders all 4 published cohorts by prandial protocol.
- MM kinetics predicts dose-proportionality up to 250 mg from first principles.
Validated against 3 published datasets. Paediatric to Swiss adult, 30–100 mg.
Coming soon: DoseTrack Full
Extended model: exercise, sleep, co-medications, and bodyweight-adjusted clearance. Personalised ADHD pharmacokinetics for clinical and research use.
Notify MeCustom PK/PD simulation
Need a simulation for a different ADHD medication, stimulant analogue, or dosing regimen? We build mechanistic PK/PD models to your specification.
Request Custom SimulationLoading report...
Team
Who We Are
Nabil Yasini
CEO & ML Research Engineer
Computational pharmacologist specialising in ADHD and stimulant pharmacology, GNN-based molecular property prediction, and mechanistic PK/PD modelling for CNS drugs.
We're growing.
Hiring across computational chemistry, ML engineering, and pharmacometrics. Domain expertise in ADHD pharmacology is a plus.
Get in touch →Research
Publications
StereoGNN-BBB: A Stereo-Aware Graph Neural Network for Blood-Brain Barrier Permeability Prediction with State-of-the-Art External Validation
Yasini N. (2025) · Independent Research
Stereo-aware GNN with 21-dimensional node features encoding R/S chirality and E/Z geometry. 0.961 AUC on B3DB external validation (7,807 compounds), outperforming 8 published methods. Focal loss improves specificity by 55%.
0.961 External AUC
7,807 compounds
GATv2 + Focal Loss
A Mechanistic Prodrug Pharmacokinetic Model for Lisdexamfetamine with First-Principles Nutrition-Aware Absorption Estimation
Yasini N. (2025) · Dis-Solved, Independent Research
Two-compartment prodrug ODE system with Michaelis-Menten kinetics. All parameters fixed to published values, no fitting to validation data. 9.1% Cmax MAPE across 6 dose-dataset combinations, 67 subjects, 3 published datasets.
9.1% Cmax MAPE
67 subjects
MM + RK4
Pricing
Start screening ADHD candidates today
Free for individual researchers. Usage-based pricing for teams.
Researcher
Free
Full access to ADHD-relevant ADMET prediction and PK/PD tools. No sign-up required.
- BBB permeability prediction
- ADMET toolkit (MAT, hERG, CYP450)
- Abuse liability scoring
- DoseTrack PK/PD simulation
- 50 predictions/day
Pro & Enterprise
Usage-based pricing
Full programmatic access and dedicated support for production screening workflows.
- REST API access
- Batch processing (CSV/SDF)
- Custom model training
- On-premise deployment
- Dedicated support & SLA
Contact
Get In Touch
Questions about the platform, custom ADHD simulations, or enterprise access?