3D structures of membrane proteins from genomic sequencing

By Thomas A. Hopf, Lucy J. Colwell, Robert P. Sheridan, Burkhard Rost, Christopher Sander, Debora Marks. 2012

517 citations recorded by the source corpus.

Summary We show that amino acid co-variation in proteins, extracted from the evolutionary sequence record, can be used to fold transmembrane proteins. We use this technique to predict previously unknown, 3D structures for 11 transmembrane proteins (with up to 14 helices) from their sequences alone. The prediction method (EVfold_membrane), applies a maximum entropy approach to infer evolutionary co-variation in pairs of sequence positions within a protein family and then generates all-atom models with the derived pairwise distance constraints. We benchmark the approach with blinded, de novo computation of known transmembrane protein structures from 23 families, demonstrating unprecedented accuracy of the method for large transmembrane proteins. We show how the method can predict oligomerization, functional sites, and conformational changes in transmembrane proteins. With the rapid rise in large-scale sequencing, more accurate and more comprehensive information on evolutionary constraints can be decoded from genetic variation, greatly expanding the repertoire of transmembrane proteins amenable to modelling by this method.

Subfields: Molecular Biology, Computational Biology & Bioinformatics

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Molecular Biology

In Computational Biology & Bioinformatics

Citations in the corpus

Coevolutionary Landscape Inference and the Context-Dependence of Mutations in Beta-Lactamase TEM-1 (2015, methodology): “In this paper we propose a variant of DCA which assigns to each mutant sequence a statistical score, which in a next step In technical terms, a mutational landscape is given as a genotype-to-phenotype mapping.”
CCMpred—fast and precise prediction of protein residue–residue contacts from correlated mutations (2014, background): “This has resulted in a boost in contact prediction accuracy, thanks to which it is now possible to reliably predict protein structures using only sequence information if enough homologous sequences are available (Hopf et al., 2012; Marks et al., 2011; Nugent and Jones, 2012).”
Protein structure prediction from sequence variation (2012, background): “1b ), protein interactions and alternative conformations 15,16,22 .”
Robust and accurate prediction of residue–residue interactions across protein interfaces using evolutionary information (2014, background): “…a global statistical model ( Thomas et al., 2008 ) to distinguish between direct and indirect couplings ( Marks et al., 2011 ; Morcos et al., 2011 ; Hopf et al., 2012 ; Nugent and Jones, 2012 ; Jones et al., 2012 ; Lapedes et al., 2012 ; Marks et al., 2012 ; Sułkowska et al., 2012 ; Kamisetty et…”
MetaPSICOV: combining coevolution methods for accurate prediction of contacts and long range hydrogen bonding in proteins (2014, background): “…that correct predicted contacts are more likely than incorrect contacts, which is an important performance threshold in terms of accurate 3D model building from predicted contacts (Hopf et al., 2012; Kosciolek and Jones, 2014; Marks et al., 2011; Michel et al., 2014; Nugent and Jones, 2012).”
Enhancing Evolutionary Couplings with Deep Convolutional Neural Networks (2017, background): “Importantly, some of the false contacts identified by DeepContact are true homodimer contacts and others may be true in another conformation, particularly given the small distances between the residues of many of the false contacts (Figures 3A and 4B) (Hopf et al., 2012; Morcos et al., 2011; Toth-Pe”
ProtTrans: Towards Cracking the Language of Life’s Code Through Self-Supervised Deep Learning and High Performance Computing (2020, background): “Third, the improvement is best when the EI is most diverse [28], [29].”
Sequence co-evolution gives 3D contacts and structures of protein complexes (2014, methodology): “work have indicated that this global probability model approach requires a minimum number of sequences in the alignment with at least 1 non-redundant sequence per residue (Marks et al., 2011; Morcos et al., 2011; Hopf et al., 2012; Jones et al., 2012; Kamisetty et al., 2013).”
Masked Language Modeling for Proteins via Linearly Scalable Long-Context Transformers (2020, neutral): “This follows recent works employing simpler statistical models that predict protein quaternary structure, protein-protein interactions and protein interaction networks from evolutionary sequence data [52, 26, 36, 3, 14].”
Large-scale determination of previously unsolved protein structures using evolutionary information (2015, background): “Hopf et al. (2012) also used co-evolution information to guide membrane protein structure prediction and function assignment; we compare to their conclusions in the two cases common to both studies.”
Improved Contact Predictions Using the Recognition of Protein Like Contact Patterns (2014, background): “It has been shown that these global contact prediction methods actually provide sufficient information for predicting the structure of proteins belonging to large protein families [7,19,20].”
Protein contact prediction by integrating joint evolutionary coupling analysis and supervised learning (2013, methodology): “Results: This article presents a group graphical lasso (GGL) method for contact prediction that integrates joint multi-family EC analysis and supervised learning to improve accuracy on proteins without many sequence homologs.”
Inverse statistical problems: from the inverse Ising problem to data science (2017, background): “This problem is addressed by an inverse approach aimed at finding the set of pairwise couplings that lead to the observed correlations or sequences [39, 55, 58, 71, 99, 138, 154, 213, 239].”