TASK	DATASET	MODEL	METRIC NAME	METRIC VALUE	GLOBAL RANK	REMOVE
Protein-Ligand Affinity Prediction	CSAR-HiQ	LightGBM	RMSE	1.725	# 2
Protein-Ligand Affinity Prediction	PDBbind	LightGBM	RMSE	1.316	# 3

Badge	Markdown
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/high-performance-of-gradient-boosting-in/protein-ligand-affinity-prediction-on-csar)](https://paperswithcode.com/sota/protein-ligand-affinity-prediction-on-csar?p=high-performance-of-gradient-boosting-in)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/high-performance-of-gradient-boosting-in/protein-ligand-affinity-prediction-on-pdbbind)](https://paperswithcode.com/sota/protein-ligand-affinity-prediction-on-pdbbind?p=high-performance-of-gradient-boosting-in)`

High Performance of Gradient Boosting in Binding Affinity Prediction

14 May 2022 · Dmitrii Gavrilev, Nurlybek Amangeldiuly, Sergei Ivanov, Evgeny Burnaev ·

Prediction of protein-ligand (PL) binding affinity remains the key to drug discovery. Popular approaches in recent years involve graph neural networks (GNNs), which are used to learn the topology and geometry of PL complexes. However, GNNs are computationally heavy and have poor scalability to graph sizes. On the other hand, traditional machine learning (ML) approaches, such as gradient-boosted decision trees (GBDTs), are lightweight yet extremely efficient for tabular data. We propose to use PL interaction features along with PL graph-level features in GBDT. We show that this combination outperforms the existing solutions.

PDF Abstract