SAPIAN™
THE UNIVERSAL INTERACTION ENGINE

Proteome-scale interaction prediction to unlock novel, patentable hits - faster and safer.

AN ACCELERATOR

Sapian is Kantify's AI-driven Dry Lab platform for hit discovery. It predicts protein-ligand interactions across the entire eukaryotic proteome, enabling discovery teams to identify novel, IP-free chemotypes - before committing a single wet lab experiment.

Instead of computer-based docking or lab-based high-throughput screening, Sapian uses probabilistic inference to identify novel or first in class chemotypes for both druggable and traditionally "undruggable" targets without requiring crystal structures.

Petri dish research

Because interactions are predicted globally, Sapian enables:

  • Early selectivity assessment
  • Built-in safety signals
  • Rapid prioritization of viable scaffolds before synthesis

The result is faster hit discovery, stronger IP, and lower early-stage attrition, all through a single dry lab engine applicable across therapeutic areas.

Human biology

FROM DOCKING TO DISCOVERY INTELLIGENCE

Traditional hit discovery is constrained by:

  • Limited structural availability and reliability
  • Local pocket modeling
  • Slow, stochastic wet-lab screening

AI-powered methods have yet to master the complexity of biochemical and phenotypic reactions. This gap is most evident with undruggable targets, where traditional predictive models consistently fall short.

Sapian moves beyond these limitations by treating drug discovery as a large-scale interaction inference problem. Rather than modeling a single binding site, Sapian predicts the probability of interaction between any molecule and any protein, generating a global view of biological compatibility, identifying multiple possible novel chemotypes for a target of interest, and flagging potential off-target risks early.

Nicolas Maignan

We map the likelihood of a biological handshake, regardless of structures. This opens multiple opportunities.

Nicolas Maignan, COO, Kantify

CRACKING THE "UNDRUGGABLE"

Sapian is built as a scalable factory for small molecule discovery. Sapian's speed is achieved by an interaction model with strong generalization power across unseen chemotypes and proteins.

As a result, Sapian can infer binding likelihoods in chemical space orders of magnitude larger than any experimentally screened or structurally characterized dataset, even in the absence of structural information.

Approach

Novelty-led discovery to identify untapped chemical space.
Selectivity optimization guided by proteome-wide predictions.

Outcome

  • First-in-class starting points for targets previously written off.
  • Orders-of-magnitude faster hit identification at a fraction of traditional HTS cost.
Discovery factory

SAFETY-BY-DESIGN

Because Sapian evaluates interactions globally, safety is integrated from the very beginning.

  • Early hERG interaction prediction at the hit identification stage (responsible for the withdrawal of several marketed drugs);
  • Removal of potential patented structures;
  • Chemotypes with high cardiotoxicity risk are deprioritized before synthesis;
  • ADMETox endpoint prediction (BBB, DILI, ...);
  • Proteome-wide visibility enables early detection of off-target liabilities.

Outcome

Medicinal chemistry resources are focused only on scaffolds that are both biologically relevant and toxicologically viable.

Safety by design

WHY CHOOSE SAPIAN

IP Frontiering

Identify novel, unclaimed chemotypes in dark chemical space

→ Stronger patents and freedom to operate

Universal Scope

One engine pre-trained to tackle novel targets regardless of available data

→ Rapid target pivoting without retooling

Dry Lab Velocity

Replace months of wet-lab screening with rapid inference

→ 100× more hits per euro or dollar invested

Drugging the Undruggable

Generate hits where docking and physics fail

→ Access to first-in-class opportunities

Dr. Rubal Ravinder

Sapian is a versatile dry factory, which finds novel chemotypes amongst billions, even before wet-lab commitment or before IP is crowded.

Dr. Rubal Ravinder,
Biochemistry Machine Learning Engineer, Kantify

HOW SAPIAN COMPARES WITH HIGH-THROUGHPUT SCREENING

>10,000× acceleration in hit discovery

Typical HTS throughput (~100,000 compounds/day), experimentally screening a 10-billion-molecule library would take ~270 years.

Minimum 100-fold improvement in performance

Hit rate in HTS: 0.1%. Hit rate with Sapian: between 10% and 96% (target dependent, difficulty is systematically evaluated ex ante).

10,000 times more molecules evaluated

Average campaign size in big pharma: 1 million compounds. Average campaign size with Sapian: 10 billion compounds.

Contact us for a non-binding and confidential feasibility analysis for your therapeutic target before any pilot.

PARTNERSHIPS

Get in touch !