Related Experiment Video
Updated: Jun 3, 2025
10:21
Author Spotlight: Streamlining Protein Target Prediction and Validation via Molecular Docking and CETSA
Published on: February 23, 2024
2.3K
Coverage bias in small molecule machine learning
Fleming Kretschmer1, Jan Seipp2, Marcus Ludwig1,3
1Chair for Bioinformatics, Institute for Computer Science, Friedrich Schiller University Jena, Jena, Germany.
Nature Communications
|January 9, 2025
Summary
Machine learning models for small molecules often lack coverage of biomolecular structures. This study introduces a new method to assess dataset coverage, improving model performance by guiding future data creation.
Area of Science:
- Computational chemistry
- cheminformatics
- machine learning
Background:
- Small molecule machine learning predicts properties from structures for applications like toxicity and drug discovery.
- End-to-end models are trending, but often overlook the domain of applicability and data coverage bias.
Purpose of the Study:
- To investigate the coverage of biomolecular structure space in large-scale datasets used for machine learning.
- To develop methods for assessing dataset representativeness and guiding future dataset creation.
Main Methods:
- Proposed a novel distance measure based on the Maximum Common Edge Subgraph (MCES) problem to quantify chemical similarity.
- Developed an efficient computational approach combining Integer Linear Programming and heuristic bounds to solve the MCES problem.
Main Results:
- Found that many widely-used datasets exhibit non-uniform coverage of biomolecular structures.
- This lack of uniform coverage limits the predictive power of machine learning models trained on these datasets.
Conclusions:
- Dataset coverage is a critical, often overlooked, factor in small molecule machine learning.
- The proposed MCES-based distance and divergence assessment methods can guide the creation of more representative datasets, enhancing model performance.
Related Concept Videos
01:23Mechanistic Models: Compartment Models in Individual and Population Analysis
27
Mechanistic models are utilized in individual analysis using single-source data, but imperfections arise due to data collection errors, preventing perfect prediction of observed data. The mathematical equation involves known values (Xi), observed concentrations (Ci), measurement errors (εi), model parameters (ϕj), and the related function (ƒi) for i number of values. Different least-squares metrics quantify differences between predicted and observed values. The ordinary least...
27
01:26Drug Discovery: Overview
7.4K
Drug discovery is a multifaceted process involving extensive screening, testing, and optimization of lead compounds to identify potential new drugs for therapeutic use. It combines several approaches, including screening large numbers of natural products, chemical modification of known active molecules, identification of new drug targets, and rational design based on biological mechanisms and drug-receptor structure. These approaches are carried out in both academic research laboratories and...
7.4K
01:26Pharmacokinetic Models: Comparison and Selection Criterion
38
Physiological and compartmental models are valuable tools used in studying biological systems. These models rely on differential equations to maintain mass balance within the system, ensuring an accurate representation of the dynamic processes at play.
Physiological models take a detailed approach by considering specific molecular processes. They can predict drug distribution, metabolism, and elimination changes, providing a comprehensive understanding of how drugs interact with the body.
Physiological models take a detailed approach by considering specific molecular processes. They can predict drug distribution, metabolism, and elimination changes, providing a comprehensive understanding of how drugs interact with the body.
38
02:40Ligand Binding Sites
12.7K
Proteins are dynamic macromolecules that carry out a wide variety of essential processes; however, the activities of most proteins depend on their interactions with other molecules or ions, known as ligands.
Protein-ligand interactions are quite specific; even though numerous potential ligands surround a cellular protein at any given time, only a particular ligand can bind to that protein. Moreover, a ligand binds only to a dedicated area on the surface of the protein, known as the...
Protein-ligand interactions are quite specific; even though numerous potential ligands surround a cellular protein at any given time, only a particular ligand can bind to that protein. Moreover, a ligand binds only to a dedicated area on the surface of the protein, known as the...
12.7K
01:49Conserved Binding Sites
4.2K
Many proteins’ biological role depends on their interactions with their ligands, small molecules that bind to specific locations on the protein known as ligand-binding sites. Ligand-binding sites are often conserved among homologous proteins as these sites are critical for protein function.
Binding sites are often located in large pockets, and if their location on a protein’s surface is unknown, it can be predicted using various approaches. The energetic method computationally...
Binding sites are often located in large pockets, and if their location on a protein’s surface is unknown, it can be predicted using various approaches. The energetic method computationally...
4.2K
02:18The Equilibrium Binding Constant and Binding Strength
12.8K
The equilibrium binding constant (Kb) quantifies the strength of a protein-ligand interaction. Kb can be calculated as follows when the reaction is at equilibrium:
12.8K