Deep learning allows genome-scale prediction of Michaelis constants from structural features

The Michaelis constant KM describes the affinity of an enzyme for a specific substrate and is a central parameter in studies of enzyme kinetics and cellular physiology. As measurements of KM are often difficult and time-consuming, experimental estimates exist for only a minority of enzyme-substrate...

Full description

Saved in:
Bibliographic Details
Published in:PLoS biology 2021-10, Vol.19 (10), p.e3001402-e3001402
Main Authors: Kroll, Alexander, Engqvist, Martin K M, Heckmann, David, Lercher, Martin J
Format: Article
Language:English
Subjects:
Amino acid sequence
Amino acids
Artificial intelligence
Binding sites
Biology and Life Sciences
cell metabolism
Chemical fingerprinting
Computer and Information Sciences
Databases, Genetic
Deep Learning
E coli
enzyme substrate
Enzymes
Enzymes - metabolism
Genome
Genomes
Graph neural networks
Kinetics
Learning algorithms
Metabolism
Metabolites
Metabolomics
Methods and Resources
Michaelis constant
Models, Biological
molecular fingerprinting
Neural networks
Neural Networks, Computer
Nucleotide sequence
Organisms
Parameterization
Physical Sciences
Physiology
Predictions
Proteins
Research and Analysis Methods
Substrate Specificity
Substrates
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The Michaelis constant KM describes the affinity of an enzyme for a specific substrate and is a central parameter in studies of enzyme kinetics and cellular physiology. As measurements of KM are often difficult and time-consuming, experimental estimates exist for only a minority of enzyme-substrate combinations even in model organisms. Here, we build and train an organism-independent model that successfully predicts KM values for natural enzyme-substrate combinations using machine and deep learning methods. Predictions are based on a task-specific molecular fingerprint of the substrate, generated using a graph neural network, and on a deep numerical representation of the enzyme's amino acid sequence. We provide genome-scale KM predictions for 47 model organisms, which can be used to approximately relate metabolite concentrations to cellular physiology and to aid in the parameterization of kinetic models of cellular metabolism.
ISSN:1545-7885
1544-9173
1545-7885
DOI:10.1371/journal.pbio.3001402