Peptide & Aptamer Design Software

LLead AI Diagnostics workspacePlatform demo records

Virus target
Binding region
Peptide rules
AI pipeline
05Results

RUN-001 / CANDIDATE REVIEW

From a sequence to a next step.

2 shortlisted

Candidate / sequencePrototype scoreShortlist

CAND-001 / WHY IT IS HERE

HBGA pocket

Matches the selected constraints in the platform’s demo configuration.

Length10 residuesSourceRUN-001Next stepScientist review

Select a candidate. Build an example shortlist.Demo sequences and scores · not laboratory evidence

Supported workflowPeptide & Aptamer Design

Why teams come
to us for this.

Binding-molecule sequence space is too large to test or simulate — and a model score must never pass as lab proof.

Norovirus: 6,037 candidates simulated to screen a 603,750-peptide library
Classifier performance of 0.84–0.91 AUC-ROC
Lead docking score of −75.4 kcal/mol vs a −51.1 training-set mean

DESIGNED AROUND THE TARGET

In silico peptide screening,
from library to shortlist.

Constrain the search. Simulate a sample.
Carry the evidence through to the next experiment.

01

Target-specific library

A peptide or aptamer library built around your target and use, such as a biosensor ligand.

Select a region to explore the configuration

01 / TARGET & BINDING REGION4OP7

Choose where to look.

Norovirus GII.4
VP1 capsid P domain

02 / PEPTIDE RULESExample configuration

GGG · prefixVariable coreCY · suffix

Length 9–16 aaTop candidates 10Diversity Balanced

Focus region: HBGA pocket

An example peptide configuration from the platform’s screening flow.

02

Representative simulation

An affordable sample simulated to create real training labels.

Compare the simulated sample with model coverage

NOROVIRUS RESEARCH EXAMPLE1%

simulated to label the sample

Simulation-labelledModel-scoredEach cell ≈ 1% of the library

SIMULATION → TRAINING

Small sample.
Real simulation labels.

6,037Candidates simulated603,750Peptides in the full library

Select a representative subset
Simulate the selected candidates
Train on the resulting labels

The orange cell represents the subset used to generate training labels.

03

Model scoring at larger scale

The model scores candidates simulation could never reach.

Explore the training evidence and scored library

MODEL / SCORING RUNPrediction

Simulation labelsTrained modelLibrary scores

Prototype candidateModel score

CAND-00196

CAND-00294

CAND-00391

Platform demo scores shown to illustrate the review interface.

REPORTED CLASSIFIER PERFORMANCE0.84–0.91 AUC-ROC

Norovirus research example · prediction is not lab proof

04

Independent-method consensus

Top candidates cross-checked by independent methods before advancing.

Select a candidate to compare the methods

INDEPENDENT-METHOD COMPARISON

Does the evidence agree?

Illustrative checks

CandidatePrimary rankingIndependent check

Example cross-check outcomes, not measured results.

CAND-003 / NEXT DECISIONCheck disagreement

Different rankings. A visible decision.

The independent check does not support the same conclusion. Keep the difference visible for a scientist to investigate.

SR

Scientific reviewExperimental confirmation still required

Built for

Research leadershipDiagnostic R&DComputational scientists

The rest of the Research & Development

Supported workflow

Therapeutic

A shortlist you can defend.

Supported workflow

ADMET

Catch liabilities early.

Want Diagnostics running on your data?

Tell us the process or prediction you need. We'll map it to a module, a deployment and an honest timeline.

Bring your challenge.
Let’s find the right next step.

Request a Free Data Assessment