Protein function predictions articles from across Nature Portfolio

Although it is possible to use deep learning models to predict static protein conformations from sequencing data, proteins are not static biochemical artefacts. ItsFlexible is a graph-based deep learning tool that is trained on a new dataset of experimentally captured protein motif conformations to classify the dynamic characteristics of proteins.

A new approach sheds light on the biological features learned by protein language models, promising greater interpretability for unsupervised sequence learning.

Signal peptide sequences appear in secreted proteins. Here, authors use these sequences in a proteomic search method called “signal strapping” to discover metalloproteins, which have an N-terminal histidine that chelates a transition metal ion.

FUGAsseM predicts protein function in microbiomes using coexpression patterns from metatranscriptomes and diverse community-wide data.

Bacteriophages (the viruses that infect bacteria) play key roles in microbial communities, but the functions of most of their genes remain unknown. Here, Boulay et al. present a machine-learning classifier that uses protein language models to assign functions to bacteriophage proteins more accurately than existing approaches.

In this work the authors present an approach called ProFlex which compresses flexibility information from protein structures via normal mode analysis, enabling scalable insights into protein dynamics and function through a novel structural alphabet.

Although it is possible to use deep learning models to predict static protein conformations from sequencing data, proteins are not static biochemical artefacts. ItsFlexible is a graph-based deep learning tool that is trained on a new dataset of experimentally captured protein motif conformations to classify the dynamic characteristics of proteins.

A new approach sheds light on the biological features learned by protein language models, promising greater interpretability for unsupervised sequence learning.

Thoughtfully designed negative training datasets may hold the key to more robust machine learning models. Ursu et al. reveal how negative training data composition shapes antibody prediction models and their generalizability. Sometimes, the best way to get better is to train harder.

A recent study introduces a series of approaches that predict protein fitness and stability after the introduction of mutations. The work focuses on combining different data and pre-training to overcome data scarcity.