PreditX® - AI Drug Discovery Workspace

PreditX® workflow

Train models, screen molecules, and review ADMET-informed results.

PreditX® is an asynchronous discovery workspace. Start by entering a biological target, select the molecular representations and machine learning models you want to train, then use completed models to screen molecules and review ranked outputs with ADMET context.

Start model training View training jobs Open screening jobs

How the app works

Train Enter target, choose descriptors, choose models, and submit an async job.

Screen Use trained models to screen uploaded molecules, pasted SMILES, or the internal database.

Review Open rankings, consensus predictions, reports, CSV files, and ADMET workbooks.

Your main workflow

Use these three steps in order. You can leave the app while training or screening runs; jobs continue asynchronously and can be checked later.

Training

Train models

Create target-specific machine learning models by entering a target, selecting molecular representations, and choosing the algorithms to train.

Enter the biological target.

Select fingerprints, Mordred descriptors, or both.

Choose the ML models and submit the training job.

Start model training

Screening

Screen molecules

After models are trained, screen molecules using a completed training job. You can upload your own database, paste SMILES directly, or use the PreditX® internal database.

Upload a molecule file for screening.

Paste up to 10,000 SMILES directly on screen.

Use the internal purchasable molecule database.

Choose trained models

Results

Review results & ADMET

Review ranked molecules, prediction confidence, consensus signals, and the ADMET profile generated for the screened molecules.

Open HTML reports and CSV outputs.

Review ADMET and developability context.

Prioritize candidates for expert review or follow-up.

View results & ADMET

Asynchronous workflow

You do not need to keep the page open.

Training and screening jobs run asynchronously. After submitting a job, you can return to the dashboard or screening job list to check status, open completed outputs, and download reports.

Use Train models to start a target-specific model training job.

Use All training jobs to monitor model training progress.

Use Screening jobs to review completed screening outputs and ADMET files.

Scientific note

Decision support, not replacement.

PreditX® outputs should support scientific prioritization. Predictions, model confidence, consensus agreement, and ADMET signals should be reviewed together with medicinal chemistry judgement and experimental feasibility.

The platform helps reduce the search space and focus follow-up decisions, but it does not replace expert interpretation or experimental validation.