Chemsift

Smarter libraries.
Faster hits. Better drugs.

ChemSift is an AI-powered platform
that selects and designs chemical libraries using
bioactivity-informed descriptors.

Screen fewer, better compounds and find more hits.

What ChemSift does

Starts from any compound library and represents each molecule with bioactivity-informed signatures, not just chemical structure

Placeholder
How it works

A ready-to-screen index of compound libraries for efficient navigation of the chemical space.

  1. Feature & Spaces

    Upload your compounds
    or pick a pre-registered library
    such as a diversity, drug-like.
  2. Feature & Spaces

    Upload your compounds
    or pick a pre-registered library
    such as a diversity, drug-like.
  3. Feature & Spaces

    Upload your compounds
    or pick a pre-registered library
    such as a diversity, drug-like.
  4. Feature & Spaces

    Upload your compounds
    or pick a pre-registered library
    such as a diversity, drug-like.
Use cases

Select the best set
of compounds
for your experiments

1

Trim a large library

Select a representative, non-redundant subset that fits your budget.
2

Expand from your hits

Pull similar or dissimilar compounds from registered libraries.
3

Pick by property

Enrich for a target or class, or deplete unwanted properties
Why ChemSift

Built on top of advanced, data-driven molecular representations

Most selection tools compare molecules by structure alone. ChemSift compares them by biology. It is built on the Chemical Checker and Bioactivity Signaturizers — validated, peer-reviewed, open-source methods that encode how compounds behave across biological assays.
Our team

Cheminformatics and systems biology expertise

Martino Bertoni — CTO.

Lead developer of the Bioactivity Signaturizers; senior data scientist. Backed by IRB Barcelona and ICREA.

Patrick Aloy — CEO

ICREA Research Professor; computational biology and AI-driven bioactivity descriptors.

Miquel Duran-Frigola — CSO.

Lead developer of the Chemical Checker; ML/AI models and cloud deployment.