AptaPilot · Aptamer discovery, done right · Waterloo, Canada

Turn stalled SELEX data into candidates worth testing.

Your pool enriched, but abundance ranking didn't give you a binder. We re-analyse the sequencing under your real buffer conditions and return a ranked, structure-aware shortlist to synthesise and test. When you need a validated aptamer with a measured Kd, our Discovery service runs the selection with the partner lab.

Request a free assessment →See a sample deliverable
Rescue→Discover→Characterize→License
Evidence

Numbers with denominators, and what they don't mean.

Rescues
6 of 22
Pools that had enriched but yielded no binder by abundance ranking, re-analysed by us; six produced a consistent family with a synthesis-ready shortlist. Internal lab records, not yet peer-reviewed. The other 16: no candidate found — which is not proof that no binder exists.
Family recovery
F1 0.89
On a public doped-SELEX library with real ground truth, the structure-aware classifier recovered the spiked families at F1 0.89 (previous DBSCAN method: 0.72, needing 126 clusters). A family-recovery metric — it is not a binding-prediction success rate.
First external project
3–4 weeks
Two HT-SELEX datasets (about 1.7 GB of raw FASTQ) from a European university virology group, delivered as a ranked shortlist and analysis report under a signed service agreement. Binding was not assayed at delivery.

Every benchmark reruns from one command. Setup, comparison tools, P-values and known limits are on the Validation & Methods → page.

Backed by

Non‑dilutive support, and the programmes behind it.

Funded without giving up equity: an NSERC-funded stipend through the Labs4 Market to Lab Program for the Winter 2027 cohort, and a Mitacs Business Strategy Internship approved in September 2026. The work runs with Boundless Accelerator as partner incubator and Humber Polytechnic as academic host, and its founder was selected, from applicants worldwide, for a 2026 Chunhui Outstanding Achievement in the academic track run by China’s Ministry of Education.

Prof. Juewen LiuProf. Juewen LiuScientific advisor · Tier‑1 Canada Research Chair, University of Waterloo NSERCFunds the Market to Lab stipendVia Labs4 · Winter 2027 cohort MitacsBusiness Strategy InternshipFunded · approved Sept 2026
春晖CHUNHUIOutstanding Achievement 2026China Ministry of Education · academic track, selected worldwide
Labs4Market to Lab ProgramWinter 2027 cohort · also TRL Level‑Up Boundless AcceleratorPartner incubatorCommercialisation mentorship & introductions Humber PolytechnicAcademic hostOffice of Research & Innovation Lab2MarketValidate FoundationsCompleted Sept 2026

Funding is awarded by the programmes named above; amounts are disclosed on request.

Taking projectsRescueDiscoveryCharacterize
Small-molecule · capture-SELEX
What we do

Three services. Only one ends in a validated binder.

Rescue and Characterize are computational: they turn data you already have into candidates and reports. Discovery adds the wet lab — a selection run with our academic partner laboratory and a binder measured before hand-over. We say which is which on every page.

Start here
01

Rescue

Your selection stalled. Send us the pool.

SELEX enriched, but nothing bound — or truncation killed it. We re-analyse your existing sequencing under real buffer conditions, find the families that matter, and hand back the candidates worth synthesising.

Deliver: a ranked, structure-aware shortlist with truncation points — candidates for synthesis and testing, not yet a validated binder.
Labs & biotech with a poolView →
02

Discovery

No bench? We take it from target to binder.

Bring a target. We design the selection strategy, run it with our academic partner laboratory, drive the analysis every round, and validate the winner — specialising in the hard cases: small molecules and capture-SELEX.

Deliver: a validated aptamer with measured Kd. Milestone-priced, tiered by the affinity reached.
Early biotech · diagnosticsView →
03

Characterize

You have the aptamer. We profile it.

Buffer-realistic folding, G-quadruplex detection, binding-family placement, and clean Kd fitting on candidates you already have — the report you'd struggle to generate yourself.

Deliver: a structure & binding characterisation report, add-on or standalone. Computational, from your data.
Anyone with candidatesView →

Prefer to run it yourself? The same engine is available as software → — free for academics, licensed for companies and CROs.

A real project

Three targets, twenty-three days, and a report you can audit.

5.38 M
reads across three targets and six sequenced rounds, from a research group whose selections had enriched but not resolved. Family sheets went back the same day, the first report on day 4, the combined report on day 23. Validation status: a priority order for synthesis and testing — binding was not assayed at delivery, and the report says so.
Case study: three targets, twenty-three days
Case study · De-identified

How it ran, step by step

Nine analysis steps that each changed a number, two independent pipelines, and what the deliverable deliberately does not claim.

 What a rescue report contains
Sample deliverable · What shipped

What you actually receive

Combined report, per-target annexes, a chain-by-chain appendix and a machine-readable table — with stated thresholds and input checksums.
See it work

From a pool to a shortlist, in one pass.

Full pipeline demo~5 min · 4K version on YouTube ↗
Run it yourself

Four tools, one engine.

The software behind every service — buffer-realistic folding, structure clustering and clean Kd fitting. Free for academic labs, licensed for companies and CROs. Registered with the Canadian Intellectual Property Office (five works, 2026).

Aptamer Analysis

SELEX read trimming, family clustering, and reports.

→

AptaFold

Real-buffer folding with native G-quadruplex detection.

→

Kd Fitting

Quadratic-model Kd fitting, right in the browser.

→

Structural Family

Cluster by loop structure, not full-length sequence.

→
Academic labs · free

Get the Launcher

One Windows app installs and updates the desktop tools; Kd Fitting runs in any browser. Three of the four modules are in daily use in the partner lab. Account-based licensing is rolling out module by module — single sign-on across all four is planned for the first half of 2027.

Download the Launcher →
Companies & CROs · licensed

Seats, site licences, OEM

Commercial seats through the same Launcher, per-site licences with calibration updates, or the engine embedded in your own service. Commercial rights on service results are negotiated per project.

For companies & CROs →
How we work

Simple terms. Shared risk.

Free assessment
One pool
We check whether your data can be analysed, tell you how many families we see, and preview up to three candidates. The full ranked shortlist, truncation map and report are the paid analysis.
Confidentiality
NDA first
Signed before any data moves — ours or yours. Data is used only for your project, never shared with third parties, and deleted on request when the project closes. The partner lab sees it only for wet-lab work you commission.
Pricing
Fixed · staged
Rescue screen: a fixed fee per pool. Full rescue in three stages — a base fee that covers cost, a delivery fee, and a success fee only after your own binding assay confirms the Kd. Discovery is milestone-based.
Software
Free for academics
The same engine ships as desktop software — free for academic labs, licensed for companies and CROs. A seat is priced below one failed synthesis round.

Have a pool that didn't give you a binder?

Request a free assessment — one pool, no charge, NDA first.
Request a free assessment →[email protected]
AptaPilot — aptamer discovery, done right · Waterloo, CanadaFounder · Xiaohan Zhang