OPAL+: Length‐Specific MoRF Prediction in Intrinsically Disordered Protein Sequences

Abstract

Intrinsically disordered proteins (IDPs) contain long unstructured regions, which play an important role in their function. These intrinsically disordered regions (IDRs) participate in binding events through regions called molecular recognition features (MoRFs). Computational prediction of MoRFs helps identify the potentially functional regions in IDRs. In this study, OPAL+, a novel MoRF predictor, is presented. OPAL+ uses separate models to predict MoRFs of varying lengths along with incorporating the hidden Markov model (HMM) profiles and physicochemical properties of MoRFs and their flanking regions. Together, these features help OPAL+ achieve a marginal performance improvement of 0.4–0.7% over its predecessor for diverse MoRF test sets. This performance improvement comes at the expense of increased run time as a result of the requirement of HMM profiles. OPAL+ is available for download at https://github.com/roneshsharma/OPAL-plus/wiki/OPAL-plus-Download.

Computational Prediction of MoRFs, Short Disorder-to-order Transitioning Protein Binding Regions

“…For instance, the newest OPAL+ method uses MoRF predictions generated by MoRFpred-plus and MoRFchibi as inputs to its SVM model. Correspondingly, the OPAL+ model is shown to be more accurate than these two input predictors on all test datasets [ 100 ].…”

Torches, Candles, Lamps, Lanterns, Flashlights, Spotlights, Night Vision Goggles… You Need Them All to See in Darkness

“…[12] Sharma et al describe a computational tool OPAL+ for identification in IDPs/IDPRs-specific disorder-based binding regions, known as molecular recognition features (MoRFs). [13] This tool represents an improved MoRF predictor that uses multiple support vector machine (SVM) models, each trained using MoRFs of different lengths, utilizes evolutionary information of the ID-PRs, as well as incorporates the hidden Markov model (HMM) profiles and utilizes physicochemical properties of MoRFs and their flanking regions. [13] Erdős et al [14] represent the results of the utilization of IUPred2A, which is a recently developed sequence-based predictor of regions that are likely to undergo disorder-to-order transitions induced by redox changes, [15] to estimate the abundance of the redox-sensitive conditionally disordered regions in various proteomes.…”

“…[13] This tool represents an improved MoRF predictor that uses multiple support vector machine (SVM) models, each trained using MoRFs of different lengths, utilizes evolutionary information of the ID-PRs, as well as incorporates the hidden Markov model (HMM) profiles and utilizes physicochemical properties of MoRFs and their flanking regions. [13] Erdős et al [14] represent the results of the utilization of IUPred2A, which is a recently developed sequence-based predictor of regions that are likely to undergo disorder-to-order transitions induced by redox changes, [15] to estimate the abundance of the redox-sensitive conditionally disordered regions in various proteomes. [14] The authors show that regions characterized by the redox-sensitive conditional disorder are very common and have multiple important functions, often related to the regulation of a wide spectrum of biological processes, and that mutations in such regions can be related to the pathogenesis of several diseases.…”

“…Sharma et al. describe a computational tool OPAL+ for identification in IDPs/IDPRs‐specific disorder‐based binding regions, known as molecular recognition features (MoRFs) . This tool represents an improved MoRF predictor that uses multiple support vector machine (SVM) models, each trained using MoRFs of different lengths, utilizes evolutionary information of the IDPRs, as well as incorporates the hidden Markov model (HMM) profiles and utilizes physicochemical properties of MoRFs and their flanking regions …”

HseSUMO: Sumoylation site prediction using half-sphere exposures of amino acids residues

“…However, a vast majority of previously proposed methods have focused on features chiefly based on amino acid occurrence patterns at the sites of interest. To explore the usefulness of structural features [28][29][30][31][32][33], we propose a novel method, Hse-SUMO that uses a combination of four different half-sphere exposure (HSE) measures, originally developed to characterize solvent exposure at particular amino acid residues [34]. We demonstrate that a combination of these features is highly promising for prediction of sumoylation sites and we were able to achieve very good levels of performance even using a relatively simple decision tree classifier (0.89 area under ROC curve for 6, 8 and 10-fold cross-validation schemes).…”