5 April 2022 in Bioinformatics
After the outstanding breakthrough of AlphaFold in predicting protein 3D models, new questions appeared and remain unanswered. The ensemble nature of proteins, for example, challenges the structural prediction methods because the models should represent a set of conformers instead of single structures. The evolutionary and structural features captured by effective deep learning techniques may unveil the information to generate several diverse conformations from a single sequence. Here, we address the performance of AlphaFold2 predictions obtained through ColabFold under this ensemble paradigm. Using a curated collection of apo–holo pairs of conformers, we found that AlphaFold2 predicts the holo form of a protein in ∼70% of the cases, being unable to reproduce the observed conformational diversity with the same error for both conformers. More importantly, we found that AlphaFold2's performance worsens with the increasing conformational diversity of the studied protein. This impairment is related to the heterogeneity in the degree of conformational diversity found between different members of the homologous family of the protein under study. Finally, we found that main-chain flexibility associated with apo–holo pairs of conformers negatively correlates with the predicted local model quality score plDDT, indicating that plDDT values in a single 3D model could be used to infer local conformational changes linked to ligand binding transitions. Data and code used in this manuscript are publicly available at https://gitlab.com/sbgunq/publications/af2confdiv-oct2021. Supplementary data are available at Bioinformatics online.
13 February 2017 in PLOS Computational Biology
MFMaría Silvina Fornasari
+1
Protein motions are a key feature to understand biological function. Recently, a large-scale analysis of protein conformational diversity showed a positively skewed distribution with a peak at 0.5 Å C-alpha root-mean-square-deviation (RMSD). To understand this distribution in terms of structure-function relationships, we studied a well curated and large dataset of ~5,000 proteins with experimentally determined conformational diversity. We searched for global behaviour patterns studying how structure-based features change among the available conformer population for each protein. This procedure allowed us to describe the RMSD distribution in terms of three main protein classes sharing given properties. The largest of these protein subsets (~60%), which we call “rigid” (average RMSD = 0.83 Å), has no disordered regions, shows low conformational diversity, the largest tunnels and smaller and buried cavities. The two additional subsets contain disordered regions, but with differential sequence composition and behaviour. Partially disordered proteins have on average 67% of their conformers with disordered regions, average RMSD = 1.1 Å, the highest number of hinges and the longest disordered regions. In contrast, malleable proteins have on average only 25% of disordered conformers and average RMSD = 1.3 Å, flexible cavities affected in size by the presence of disordered regions and show the highest diversity of cognate ligands. Proteins in each set are mostly non-homologous to each other, share no given fold class, nor functional similarity but do share features derived from their conformer population. These shared features could represent conformational mechanisms related with biological functions. Protein motions are commonly quantified measuring structural differences between conformers. The extension of these differences are called conformational diversity. These motions are essential to understand protein biology. We have found that the distribution of conformational diversity in a large dataset of proteins could be explained in terms of three sets sharing structure-based features emerging from the conformer population for each protein. The first set, which we called rigid, involve proteins showing almost no backbone movements but with important changes in tunnels. In order of increasing conformational diversity, the other sets are called partially disordered and malleable, showing disordered regions and important cavities but with different behaviour to each other. Shared features in each set could represent conformational mechanisms related with biological functions.
1 June 2015 in Current Opinion in Structural Biology
7 May 2013 in Molecular Biology and Evolution
MFMaria Silvina Fornasari
Native state of proteins is better represented by an ensemble of conformers in equilibrium than by only one structure. The extension of structural differences between conformers characterizes the conformational diversity of the protein. In this study, we found a negative correlation between conformational diversity and protein evolutionary rate. Conformational diversity was expressed as the maximum root mean square deviation (RMSD) between the available conformers in Conformational Diversity of Native State database. Evolutionary rate estimations were calculated using 16 different species compared with human sharing at least 700 orthologous proteins with known conformational diversity extension. The negative correlation found is independent of the protein expression level and comparable in magnitude and sign with the correlation between gene expression level and evolutionary rate. Our findings suggest that the structural constraints underlying protein dynamism, essential for protein function, could modulate protein divergence.
25 October 2017 in Protein Science
21 November 2016 in Bioinformatics
Motivation: MIToS is an environment for mutual information analysis and a framework for protein multiple sequence alignments (MSAs) and protein structures (PDB) management in Julia language. It integrates sequence and structural information through SIFTS, making Pfam MSAs analysis straightforward. MIToS streamlines the implementation of any measure calculated from residue contingency tables and its optimization and testing in terms of protein contact prediction. As an example, we implemented and tested a BLOSUM62-based pseudo-count strategy in mutual information analysis. Availability and Implementation: The software is totally implemented in Julia and supported for Linux, OS X and Windows. It’s freely available on GitHub under MIT license: http://mitos.leloir.org.ar. Contacts:diegozea@gmail.com or cmb@leloir.org.ar Supplementary information: Supplementary data are available at Bioinformatics online.