Arne Elofsson

Conserved mycobacteriophage protein uses structural mimicry to displace host initiation factors

The paper identifies a conserved viral protein module that displaces host RNA polymerase sigma factors to control transcription in mycobacteria.

ABAG-Rank: Improving Model Selection of AlphaFold Antibody-Antigen Complexes by Learning to Rank

ABAG-Rank improves model selection for AlphaFold3-predicted antibody-antigen complexes by learning to rank structural ensembles using geometric descriptors and confidence scores.

Evaluating Deep Learning Based Structure Prediction Methods on Antibody-Antigen Complexes

This study benchmarks deep learning methods for predicting antibody-antigen complex structures, finding that increased sampling improves accuracy but model selection remains challenging, with AlphaFold3 performing best overall.

Reliable Identification of Homodimers Using AlphaFold

The paper identifies that AlphaFold has a higher false-positive rate for homodimers than heterodimers and improves detection using a logistic regression model incorporating structural and homology data.

Blind Prediction of Complex Water and Ion Ensembles Around RNA in CASP16

This paper evaluates blind predictions of dynamic water and ion ensembles around RNA using cryo-EM densities from CASP16, finding that molecular dynamics approaches perform best.

AlphaFold3 at CASP16

This paper evaluates AlphaFold3's performance on CASP16, showing comparable results to top predictors for proteins and complexes, but with limitations in RNA and stoichiometry predictions.

Limits of deep-learning-based RNA prediction methods

This study evaluates deep-learning-based RNA structure prediction methods, finding they perform well for structured RNAs but struggle with novel folds and provide unreliable accuracy estimates.

M-Ionic: Prediction of metal ion binding sites from sequence using residue embeddings

M-Ionic predicts metal ion binding sites from protein sequences using residue embeddings, achieving high accuracy for known ions and extending predictions to additional ions.

Evaluation of AlphaFold-Multimer prediction on multi-chain protein complexes

The paper evaluates AlphaFold-Multimer's performance on multi-chain protein complexes and introduces a new interface quality score.

Towards a structurally resolved human protein interaction network

The paper uses AlphaFold2 to predict structures for 65,484 human protein interactions, identifying high confidence models and disease-related mutations.

Predicting the structure of large protein complexes using AlphaFold and Monte Carlo tree search

The paper predicts structures of large protein complexes using AlphaFold and Monte Carlo tree search, achieving accurate assemblies for some symmetric complexes.

EvoBind: in silico directed evolution of peptide binders with AlphaFold

EvoBind uses AlphaFold to design peptide binders towards a protein interface using only sequence information, showing it can distinguish binders from non-binders but not affinity differences.

Improved prediction of protein-protein interactions using AlphaFold2

AlphaFold2, fed concatenated chains with optimized organism-paired alignments and pDockQ-ranked seeds, gives acceptable models for 63% of heterodimers; the same predicted DockQ score flags interacting pairs.

The relationship between ageing and changes in the human blood and brain methylomes

This paper shows that DNA methylation changes with age vary across tissues and life stages, and that non-linear relationships outperform linear epigenetic clocks in age prediction.

Limits and potential of combined folding and docking using PconsDock.

This paper introduces PconsDock, a fold-and-dock method that predicts tertiary and quaternary structures of protein pairs without prior structural information, achieving comparable performance to traditional methods.

Improved protein docking by predicted interface residues

This paper improves protein docking by using predicted interface residues to score docking models, finding that BIPSPI outperforms existing methods.

The evolutionary history of topological variations in the CPA/AT superfamily

The paper identifies novel topologies in CPA/AT transporters using an integrated method and presents a database with structural classifications.

pyconsFold: A fast and easy tool for modelling and docking using distance predictions.

pyconsFold models and docks proteins using distance predictions, outperforming contact prediction methods and being 20 times faster than pyRosetta.

Critical Assessment of Protein Intrinsic Disorder Prediction

The paper evaluates 43 methods for predicting intrinsically disordered regions and binding regions in proteins, finding that deep learning methods outperform older physicochemical approaches but binding prediction remains challenging.

Identification of a novel fold type in CPA/AT transporters by ab-initio structure prediction

This study identifies a novel reentrant-helix-reentrant fold in CPA/AT transporters using ab-initio structure prediction, revealing structural adaptations and the utility of deep-learning-based modeling.

Accurate contact-based modelling of repeat proteins predicts the structure of Curlin and SPW repeats.

The paper predicts the structure of repeat proteins using contact-based deep learning, achieving high accuracy for five PFAM families without known structures.

Detecting Novel Sequence Signals in Targeting Peptides Using Deep Learning

The paper introduces TargetP 2.0, a deep learning method to identify N-terminal sorting signals in proteins, revealing insights into the importance of the second residue in targeting peptides.

Why do eukaryotic proteins contain more intrinsically disordered regions?

The paper investigates why eukaryotic proteins have more intrinsically disordered regions by comparing linker regions between eukaryotic and prokaryotic proteomes.

The classification of orphans is improved by combining searches in both proteomes and genomes

The paper improves orphan gene classification by combining proteome and genome searches, identifying fewer orphans in Drosophila species than previous studies.

PconsC4: fast, free, easy, and accurate contact predictions.

PconsC4 is a fast, free, and accurate contact predictor that outperforms existing methods without requiring external programs.

Improved protein model quality assessments by changing the target function

The paper improves protein model quality assessment by retraining ProQ3D with different target functions, finding contact-based methods better for identifying best models and S-score better for global quality correlation.

Methods For Estimation Of Model Accuracy In CASP12

The paper describes methods for estimating model accuracy in protein structure predictions, showing improvements in single model accuracy over previous methods.

Large-Scale Structure Prediction By Improved Contact Predictions And Model Quality Assessment

The paper introduces PconsFold2, a pipeline that improves protein structure prediction by integrating contact predictions and model quality assessment, identifying 558 Pfam families with 90% specificity.

Accurate contact predictions for thousands of protein families using PconsC3

PconsC3 improves contact prediction for small protein families, enabling accurate structure modeling for thousands of Pfam domains.

Folding of Aquaporin 1: Multiple evidence that helix 3 can shift out of the membrane core

The paper investigates the folding mechanism of Aquaporin 1, showing that helix 3 can shift out of the membrane core, supporting a more complex folding pathway.