Protein function predictions

Protein function predictions are bioinformatic methods (i.e. involving computation using software) that aim to predict the function of a protein. Protein function analyses use protein sequences and structures as inputs into software that generates experimentally testable predictions of function.

Latest Research and Reviews

Research
28 January 2023 | Open Access

The RESP AI model accelerates the identification of tight-binding antibodies

High-affinity antibodies are often identified through directed evolution but deep leaning methods hold great promise. Here the authors report RESP, a pipeline for efficient identification of high affinity antibodies, and apply this to the PD-L1 antibody Atezolizumab.
- Jonathan Parkinson
- , Ryan Hard
- & Wei Wang
Nature Communications 14, 454
Research
19 January 2023 | Open Access

Learning the protein language of proteome-wide protein-protein binding sites via explainable ensemble deep learning

An ensemble deep learning model (EDLM)-based protein-protein interaction (PPI) site identification method (EDLMPPI) accurately predicts PPIs through protein language models and outperforms state-of-the-art methods on multiple benchmark sets.
- Zilong Hou
- , Yuning Yang
- & Xiangtao Li
Communications Biology 6, 73
Research
14 January 2023 | Open Access

Optimizing variant-specific therapeutic SARS-CoV-2 decoys using deep-learning-guided molecular dynamics simulations
- Katharina Köchl
- , Tobias Schopper
- & Christian C. Gruber
Scientific Reports 13, 774
Research
16 December 2022 | Open Access

Pan-kinome of Legionella expanded by a bioinformatics survey
- Marianna Krysińska
- , Bartosz Baranowski
- & Marcin Gradowski
Scientific Reports 12, 21782
Research
02 December 2022 | Open Access

Dynamic spatiotemporal determinants modulate GPCR:G protein coupling selectivity and promiscuity

G protein coupled receptors (GPCRs) can couple to different Gα protein subfamilies either selectively or promiscuously. Here, the authors use computational approach to show that selectivity determinants are at the periphery of the GPCR—G protein interface and that promiscuous GPCRs more frequently sample the common rather than selective contacts.
- Manbir Sandhu
- , Aaron Cho
- & Nagarajan Vaidehi
Nature Communications 13, 7428
Reviews | 24 November 2022

Lynch syndrome, molecular mechanisms and variant classification
- Amanda B. Abildgaard
- , Sofie V. Nielsen
- & Rasmus Hartmann-Petersen
British Journal of Cancer, 1-9

All Research & Reviews

News and Comment

News & Views | 19 August 2022

ML helps predict enzyme turnover rates

Constraining metabolic models by enzyme capacities greatly improves genotype–phenotype predictions. Now, a method for estimating enzyme turnovers based on deep learning has been developed and used to reconstruct enzyme-constrained genome-scale metabolic models for more than 300 yeast species.
- Veda Sheersh Boorla
- , Vikas Upadhyay
- & Costas D. Maranas
Nature Catalysis 5, 655-657
News & Views | 01 June 2022

ScanNet uncovers binding motifs in protein structures with deep learning

Determining the functional properties of a protein from its structure is challenging. This study presents an interpretable deep learning model that directly learns function-bearing structural motifs from raw data, allowing accurate mapping of protein binding sites and antibody epitopes onto a protein structure.

Nature Methods 19, 660-661
Comments & Opinion | 11 January 2022

Structural biology is solved — now what?

The splendid computational success of AlphaFold and RoseTTAFold in solving the 60-year-old problem of protein folding raises an obvious question: what new avenues should structural biology explore? We propose a strong pivot toward the goal of reading mechanism and function directly from the amino acid sequence. This ambitious goal will require new data analytical tools and an extensive database of the atomic-level structural trajectories traced out on energy landscapes as proteins perform their function.
- Abbas Ourmazd
- , Keith Moffat
- & Eaton Edward Lattman
Nature Methods 19, 24-26
News & Views | 09 December 2021

Combining views for newly sequenced organisms

Newly sequenced organisms present a challenge for protein function prediction, as they lack experimental characterisation. A network-propagation approach that integrates functional network relationships with protein annotations, transferred from well-studied organisms, produces a more complete picture of the possible protein functions.
- Yingying Zhang
- , Shayne D. Wierbowski
- & Haiyuan Yu
Nature Machine Intelligence 3, 1011-1012
Research Highlights | 23 January 2017

Mining for CRISPR–Cas
- Ashley York
Nature Reviews Microbiology 15, 133
Correspondence | 29 October 2015

Avoiding abundance bias in the functional annotation of posttranslationally modified proteins
- Christian Schölz
- , David Lyon
- & Brian T Weinert
Nature Methods 12, 1003-1004