KL-Tools | Hit-TPT and FRASE

1. Input files

Protein structure (PDB) i

▢ Select a PDB file

Docked ligands (SDF) i

▢ Select an SDF file

Prediction model

Hit-TPT Post-docking enrichment score for docked protein-ligand complexes. FRASE Estimates whether the local ligand-protein environment is FRASE-like.

Output

Rank molecules by score If both models are selected, ranking uses Hit-TPT; otherwise the selected model is used.

2. Results

Waiting for input files 0 / 0

# Molecule Hit-TPT FRASE

Upload an SDF file to display molecules.

3D Molecular Viewer Upload a protein structure and docked ligand poses.

Visualization

Protein

Style Color

Ligand

Style Color

Viewer

Background

How KL-Tools works

Prepare the protein

Upload a protein structure in PDB format.

Upload docked ligands

Upload an SDF file containing the docked 3D poses of the compounds you want to evaluate.

Select the scoring model

Choose Hit-TPT, FRASE, or both. The protein-ligand graph is generated once and passed to the selected model or models.

Run the prediction

Hit-TPT provides post-docking enrichment scores. FRASE estimates whether the local ligand-protein environment is FRASE-like.

Review and download

Review the selected score or scores, inspect compounds in the 3D viewer, and download one scored SDF file.

Interpreting the score

Higher Hit-TPT scores indicate stronger model confidence for post-docking enrichment. Higher FRASE scores indicate stronger model confidence that the local environment is FRASE-like.

About KL-Tools

Hit-TPT

Hit-TPT (Hit-Triage Pretrained Transformer) is a physics-informed graph-transformer developed for post-docking enrichment in structure-based virtual screening. It evaluates docked protein-ligand complexes using local interaction graphs and 3D spatial information.

FRASE

FRASE means FRAgments in Structural Environments. FRASE-bot is the broader hit-finding workflow described in the original publication. In KL-Tools, FRASE refers to the scoring model used to estimate whether the submitted local ligand-protein environment is FRASE-like.

Developed by

Hamza Hentabli
Dmitri Kireev

Affiliation

Kireev Lab
Department of Chemistry
College of Arts and Sciences
University of Missouri
Columbia, Missouri, USA

KL-Tools usage

0Visitors

0Prediction jobs

0Molecules processed

FRASE-bot reference

An Y, Lim J, Glavatskikh M, et al. In silico fragment-based discovery of CIB1-directed anti-tumor agents by FRASE-bot. Nature Communications 15, 5564 (2024).