Iktos Engine

This repository contains the first version of Iktos Engine: a set of focused, production-scale APIs for 3D virtual screening workflows spanning both ligand-based and structure-based regimes.

For more technical depth, see the Iktos 3D Engine technical report in this repository.

APIs

• Iktos 3D Align

  • Accurate ligand alignment method, useful for 3D ligand-based (LB) virtual screening.

• Iktos 3D Dock

  • Accurate template-based docking method, useful for 3D structure-based (SB) virtual screening.

• Boltz 3D Affinity

  • Reuses the Boltz-2 affinity trunk (Passaro et al., 2025) to score binding affinity directly from a user-supplied protein–ligand complex, without the time-consuming cofolding step.

Repository layout

• 3d-align/: notebooks, client code, and benchmark assets for the Iktos 3D Align API integration.
• 3d-dock/: notebooks, client code, and benchmark assets for the Iktos 3D Dock API integration.
• boltz-3d-affinity/: notebooks, client code, and benchmark assets for the Boltz 3D Affinity API integration.
• virtual-screening-pipeline/: virtual screening pipelines combining these APIs — virtual_screening_pipeline_A.ipynb (full 3D Align → 3D Dock → Boltz 3D Affinity) and virtual_screening_pipeline_B.ipynb (3D Dock → Boltz 3D Affinity, skips Align). See virtual-screening-pipeline/README.md for how to run them, and example data under virtual-screening-pipeline/data/.
• Iktos_3D_Engine_Technical_Report.pdf : technical report with additional implementation and performance details on Iktos Engine and the APIs above. Open this PDF for deeper context beyond the README and example notebooks.

Getting started

Set up an isolated environment using any tool of your choice (venv, conda, uv, …), then install dependencies from the repo root:

# example with venv
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

Access & Contact

To request access or learn more, reach out at contact@iktos.com.

Status

This is an initial release of the Engine APIs and supporting materials; the repository will evolve as additional endpoints, examples, and benchmarking utilities are added.