Enter the Lab

Checking sign-in method...

Solubility/Stability Workspace

Target design runs, current rounds, and result triage in one workspace.

Current Project

Evolution Campaign Not connected yet

Runs in Progress 0

Fast Launch

Upload a FASTA/PDB/mmCIF file, or paste text if needed, then launch with standard defaults.

FASTA/PDB/mmCIF file or text Upload target file

Use a PDB, mmCIF, CIF, FASTA, FA, or text file for the usual Fast launch path.

No target loaded yet.

Paste FASTA/PDB/mmCIF text instead

Direct text input Use raw FASTA, PDB, or mmCIF text only when you do not have a file.

Run Notes Optional notes copied into the run prompt. Run name Optional. Leave blank to create one automatically.

Launch Options

Fast keeps ProteinMPNN at 2 sequences per backbone and rounds backbone counts up to satisfy the total output target.

Sequence Conservation Choose which sequence-conservation levels to generate. Default: 30%, 50%, 70%.

Total Output Sequences Target total ProteinMPNN designs across the selected sequence-conservation levels. PDB input uses RFD3 + BioEmu; FASTA input uses BioEmu only.

Extract Constraints from Paper (PDF)

Advanced Run Setup

Choose a workflow, attach inputs, and launch the job.

Run ID

Advanced always launches a new run. Load Request copies settings from the selected run.

Stage - State -

Input

Start with the target. Checks, notes, and literature constraints are below.

Optional notes or natural-language setup

Notes or setup instructions Leave this empty for normal step-by-step setup. Use it only to add notes or ask the system to infer settings from text.

Activity Log

Extract Literature Constraints (PDF)

Upload a paper when residues or mutation-sensitive regions should guide fixed-position choices.

Step Setup

Work through one step at a time.

Complete required inputs to enable execution.

Run status: -

AF2 Surrogate Triage

Run ProteinMPNN/SoluProt with GPU-backed ESM embedding and a fixed AF2/ColabFold acquisition budget.

Target Sequence or Structure Provide FASTA, PDB, or mmCIF input for the budget-aware triage path. Run name Optional. Leave blank to create one automatically.

Surrogate Settings

This mode disables RFD3 and BioEmu, embeds the SoluProt-passing candidate pool with the configured ESM worker, and spends AF2 only on the sampled training set plus Top K candidates.

Training AF2 samples Top K final AF2 candidates ProteinMPNN candidates per tier Default 3333 per tier gives about 9999 multi-tier designs before the shared AF2 training and Top K budgets are spent. Top K selection method RAPID labels training candidates with AF2, compares the checked models by cross-validation, then sends the best model's Top K candidates to AF2. Models to compare

Random Forest Ridge XGBoost LightGBM

RAPID compares these models on the same AF2-labelled training set. If the method is CV, the best model selects the final Top K. Optional rank ensemble

Random Forest Ridge XGBoost LightGBM

Leave empty by default. Select members only when you also want RAPID to compare a rank-averaged ensemble. Structure predictor SoluProt cutoff pLDDT cutoff RMSD cutoff

Use Evolution only when experimental measurements are available or planned. Use this Surrogate tab when the goal is a one-round compute budget reduction.

Evolution Studio

Upload a PDB/mmCIF structure and run the multi-round evolution pipeline with one click.

Target Structure Provide a PDB or mmCIF structure to evolve. Run name Optional. Leave blank to create one automatically.

Evolution Settings

Configure active-learning AF2 triage parameters for the evolution process.

Experiment-guided mode first chooses candidates to measure. After measured values are saved, the next run recommends candidates.

Candidate pool size Candidate selection mode Optimization metric Previous run with measurements

Choose a previous run with recorded experiment labels, or type a run ID manually.

Experiment Values

Record wet-lab labels for the selected run without leaving Evolution.

Assay type Result

Candidate ID Sequence ID (optional)

Objective metric Metric value

Metric unit (optional) Metric direction

Replicate ID (optional) Conditions / Notes

Recent Evolution Labels

First measurement set size Next recommendation count

Filtering

ColabFold pLDDT Cutoff ColabFold RMSD Cutoff Rosetta Relax Score/Residue Cutoff BioEmu Input-Backbone RMSD Limit RFD3 Input-Backbone RMSD Limit

Constraints

BioEmu Steering Config Fixed Positions Extra

Extract Constraints from Paper (PDF)

Workflow Studio

Run one stage at a time, adjust key parameters, and decide whether to rerun or move forward.

Execution Status

Studio run: -

Run -

Backend Stage -

Requested -

Updated -

State -

Artifact Preview

3D structures, images, or text extracts.

Select an artifact to preview it here.

Rounds Workspace

Organize projects, experiment rounds, and linked runs in one place.

Projects

Create and select campaign spaces for each target.

Rounds

Keep each design-test-learn cycle grouped under the current project.

Round Detail

Select a round to see its goal, notes, and linked runs.

Run Monitor

Track live status, scores, and recent runs.

Run ID

Detail -

Error summary

Stage -

State -

Updated -

Progress

- 0%

Scoring

Score: -

Evidence: -

Recommendation: -

Data Completeness

Select a run to show data completeness.

Auto Poll

Recent Runs

Run ID search

Show all runs (admin)

Activity Log

Artifacts

Filter outputs from the run.

Artifact Preview

3D structures, images, or text extracts.

Select an artifact to preview it here.

Agent Panel

Stage-by-stage expert consensus and recovery notes.

CATH Pipeline

Run subset pipeline jobs and inspect progress. Agent and Gemini features stay disabled here.

Managed Jobs

Job Log

Select a job to inspect its log.

Run ID

Stage - State -

Feedback 0

Experiments 0

Recommendation -

Structure Compare Studio

Run 3D/sequence diff and review WT/RFD3/BioEmu comparison in one place.

Artifact Preview

Select an artifact to preview it here.

Comparison Summary

Generate report to load comparison metrics.

Artifact File Viewer

Open PDB/FASTA/CSV and text outputs directly inside Analyze.

Artifact Preview

3D structures, images, or text extracts.

Select an artifact file to preview it in Analyze.

Feedback

Capture expert reviews and ratings.

Rating

Reasons

Artifact (optional) Stage (optional)

Comment

Recent Feedback

Experiment

Log wet-lab outcomes and metrics.

Assay Type Result

Sample ID (optional) Artifact (optional)

Candidate ID (optional) Sequence ID (optional)

Objective metric Metric value

Metric unit (optional) Metric direction

Replicate ID (optional) Metrics (JSON, optional) Conditions / Notes

Recent Experiments

Report

Generate a consolidated run summary.

Score: -

Evidence: -

Recommendation: -

Report (Markdown)

Report Review

Rate the report and provide reasons.

Rating

Reasons

Comment

Artifact Links

Run-to-Run Compare

Compare pLDDT, RMSD, Relax, SoluProt, and pass-rate deltas versus another run.

Baseline Run

Select a baseline run to load run-to-run deltas.

Hit List

Weighted ranking of final candidates with score cutoff filtering.

Score Cutoff 0 Rows

SoluProt W pLDDT W RMSD W

Identity is shown as percent and does not affect ranking/filtering.

Chart

Run hit list to render candidate charts.

Load a run to build the hit list.

Enter the Lab

Solubility/Stability Workspace

New Experiment

Load Existing Run

Analyze Results

Fast Launch

Advanced Run Setup

Input

Step Setup

AF2 Surrogate Triage

Evolution Studio

Workflow Studio

Artifact Preview

Rounds Workspace

Round Detail

Run Monitor

Artifacts

Artifact Preview

Agent Panel

CATH Pipeline

Structure Compare Studio

Artifact File Viewer

Artifact Preview

Feedback

Experiment

Report

Report Review

Run-to-Run Compare

Hit List