AI Platform

Systematically Mapping a New Universe of Chemical Space

A New Frontier of Scalable, Rational Design

Our AI platform, Nautilus, is enabling in silico predictive peptide design across diverse targets to efficiently access new functional chemical space and produce therapeutics with drug-like properties.

AI + Human Peptide Engineering

We brought together – and are qualitatively advancing – technologies to make scalable peptide drug design a reality.

  • Our novel, proprietary AI algorithms drive intelligent navigation of otherwise unexplored and inaccessible peptide chemical space to find new peptides with a desired set of properties.
  • Our purpose-built supercomputer and cutting-edge AI engine execute the algorithms far faster than otherwise possible, maximizing the accuracy of our models and, thereby, the quality of our hits.
  • Innovations we’re bringing to the lab are enabling rapid, cost-effective, scalable peptide synthesis coupled with rich data collection, to allow continual improvement over time.

This confluence is making possible a transformation in the very way that we do drug discovery, and enabling a move to AI-enabled design. Multiparameter optimization allows us to engineer a spectrum of properties into designed peptides, which radically focuses our search and accelerates our time to discovery.

Drug Design and Development of the Future

Building a peptide design platform that goes beyond merely finding hits: It improves the entire discovery process.

  • NOVEL AI ALGORITHMS

    Harmonized public and proprietary data spanning chemistry, biology and pharmacology​

    Generative AI models learn functional and phenotypic representations to create peptides in novel chemical space

    Predictive AI models perform multiparameter optimization for in silico screening and ranking of generated sequences​

  • PURPOSE-BUILT SUPER-
    COMPUTING

    Custom-designed, scalable on-prem HPC accelerates and integrates data, AI, simulation and analytics​

    Flexibly incorporate emerging technologies; higher performance at lower cost than public clouds

    Strategic partnership with Cerebras Systems: optimize advanced deep learning algorithms to speed design cycle by potentially reducing training time up to 1000x

  • PEPTIDE R&D

    Built custom-designed peptide synthesizer to enable ​design-scale synthesis

    Generating AI-quality ADME and pharmacology peptide datasets

    Identifying and incorporating new developments to improve data generation

Optimized Peptides

Diverse targets & Datasets

  • Extensible to all disease areas
  • Integrates data on indications, manufacturing, IP, etc. – not only sequence and structure data
  • Unlocks clinically validated, novel and challenging, previously undruggable targets

Peptide Design & Optimization

  • Generalizable to arbitrary peptides: linear, cyclic, nonstandard amino acids
  • Multiparameter optimization of physicochemical and pharmaceutical properties
  • Efficient exploration of peptide design space, surpassing mere libraries of hits

Synthesis & Validation

  • Efficient synthesis of optimized, ranked hits
  • Physiochemistry, pharmacology, pharmacokinetics, safety
  • Consistent, rich data collections further enhances AI-driven peptide design & optimization

INSPIRING OUR PLATFORM

Our lead program, PLG0206, was designed to create better outcomes in a challenging peptide drug class. For more information on our programs in development, please visit our Pipeline.