SPOTONE: Hot Spots on Protein Complexes with Extremely Randomized Trees via Sequence-Only Features

PPI-hotspotID for detecting protein-protein interaction hot spots from the free protein structure

Experimental detection of residues critical for protein-protein interactions (PPI) is a time-consuming, costly, and labor-intensive process. Hence, high-throughput PPI-hot spot prediction methods have been developed, but they have been validated using relatively small datasets, which may compromise their predictive reliability. Here, we introduce PPI-hotspotID, a novel method for identifying PPI-hot spots using the free protein structure, and validated it on the largest collection of experimentally confirmed PPI-hot spots to date. We explored the possibility of detecting PPI-hot spots using (i) FTMap in the PPI mode, which identifies hot spots on protein-protein interfaces from the free protein structure, and (ii) the interface residues predicted by AlphaFold-Multimer. PPI-hotspotID yielded better performance than FTMap and SPOTONE, a webserver for predicting PPI-hot spots given the protein sequence. When combined with the AlphaFold-Multimer-predicted interface residues, PPI-hotspotID yielded better performance than either method alone. Furthermore, we experimentally verified several PPI-hotspotID-predicted PPI-hot spots of eukaryotic elongation factor 2. Notably, PPI-hotspotID can reveal PPI-hot spots not obvious from complex structures, including those in indirect contact with binding partners. PPI-hotspotID serves as a valuable tool for understanding PPI mechanisms and aiding drug design. It is available as a web server (https://ppihotspotid.limlab.dnsalias.org/) and open-source code (https://github.com/wrigjz/ppihotspotid/).

Feature selection for effective prediction of SARS-COV-2 using machine learning

BACKGROUND

With rise in variants of SARS-CoV-2, it is necessary to classify the emerging SARS-CoV-2 for early detection and thereby reduce human transmission. Genomic and proteomic information have less frequently been used for classifying in a machine learning (ML) approach for detection of SARS-CoV-2.

OBJECTIVE

With this aim we used nucleoprotein and viral proteomic evolutionary information of SARS-CoV-2 along with the charge and basicity distribution of amino acids from various strains of SARS-CoV-2 to generate a disease severity model based on ML.

METHODS

All sequence and clinical data were obtained from GISAID. Proteomic level calculations were added to comprise the dataset. The training set was used for feature selection. Select K- Best feature selection method was employed which was cross validated with testing set and performance evaluated. Delong's test was also done. We also employed BIRCH clustering on SARS-CoV-2 for clustering the strains.

RESULTS

Out of six ML models four were successful in training and testing. Extra Trees algorithm generated a micro-averaged F1-score of 74.2% and a weighted averaged area under the receiver operating characteristic curve (AUC-ROC) score of 73.7% with multi-class option. The feature selection set to 5, enhanced the ROC AUC from 73.7 to 76.4%. Accuracy of the selected model of 86.9% was achieved.

CONCLUSION

The unique features identified in the ML approach was able to classify disease severity into classes and had potential for predicting risk in newer variants.

AFP-SPTS: An Accurate Prediction of Antifreeze Proteins Using Sequential and Pseudo-Tri-Slicing Evolutionary Features with an Extremely Randomized Tree

The development of intracellular ice in the bodies of cold-blooded living organisms may cause them to die. These species yield antifreeze proteins (AFPs) to live in subzero temperature environments. Additionally, AFPs are implemented in biotechnological, industrial, agricultural, and medical fields. Machine learning-based predictors were presented for AFP identification. However, more accurate predictors are still highly desirable for boosting the AFP prediction. This work presents a novel approach, named AFP-SPTS, for the correct prediction of AFPs. We explored the discriminative features with four schemes, namely, dipeptide deviation from the expected mean (DDE), reduced amino acid alphabet (RAAA), grouped dipeptide composition (GDPC), and a novel representative method, called pseudo-position-specific scoring matrix tri-slicing (PseTS-PSSM). Considering the advantages of ensemble learning strategy, we fused each feature vector into different combinations and trained the models with five machine learning algorithms, i.e., multilayer perceptron (MLP), extremely randomized tree (ERT), decision tree (DT), random forest (RF), and AdaBoost. Among all models, PseTS-PSSM + RAAA with an extremely randomized tree attained the best outcomes. The proposed predictor (AFP-SPTS) boosted the accuracies of AFPs in the literature by 1.82 and 4.1%.

SYNPRED: prediction of drug combination effects in cancer using different synergy metrics and ensemble learning

BACKGROUND

In cancer research, high-throughput screening technologies produce large amounts of multiomics data from different populations and cell types. However, analysis of such data encounters difficulties due to disease heterogeneity, further exacerbated by human biological complexity and genomic variability. The specific profile of cancer as a disease (or, more realistically, a set of diseases) urges the development of approaches that maximize the effect while minimizing the dosage of drugs. Now is the time to redefine the approach to drug discovery, bringing an artificial intelligence (AI)-powered informational view that integrates the relevant scientific fields and explores new territories.

RESULTS

Here, we show SYNPRED, an interdisciplinary approach that leverages specifically designed ensembles of AI algorithms, as well as links omics and biophysical traits to predict anticancer drug synergy. It uses 5 reference models (Bliss, Highest Single Agent, Loewe, Zero Interaction Potency, and Combination Sensitivity Score), which, coupled with AI algorithms, allowed us to attain the ones with the best predictive performance and pinpoint the most appropriate reference model for synergy prediction, often overlooked in similar studies. By using an independent test set, SYNPRED exhibits state-of-the-art performance metrics either in the classification (accuracy, 0.85; precision, 0.91; recall, 0.90; area under the receiver operating characteristic, 0.80; and F1-score, 0.91) or in the regression models, mainly when using the Combination Sensitivity Score synergy reference model (root mean square error, 11.07; mean squared error, 122.61; Pearson, 0.86; mean absolute error, 7.43; Spearman, 0.87). Moreover, data interpretability was achieved by deploying the most current and robust feature importance approaches. A simple web-based application was constructed, allowing easy access by nonexpert researchers.

CONCLUSIONS

The performance of SYNPRED rivals that of the existing methods that tackle the same problem, yielding unbiased results trained with one of the most comprehensive datasets available (NCI ALMANAC). The leveraging of different reference models allowed deeper insights into which of them can be more appropriately used for synergy prediction. The Combination Sensitivity Score clearly stood out with improved performance among the full scope of surveyed approaches and synergy reference models. Furthermore, SYNPRED takes a particular focus on data interpretability, which has been in the spotlight lately when using the most advanced AI techniques.

SARS-CoV-2 Membrane Protein: From Genomic Data to Structural New Insights

Severe Acute Respiratory Syndrome CoronaVirus-2 (SARS-CoV-2) is composed of four structural proteins and several accessory non-structural proteins. SARS-CoV-2's most abundant structural protein, Membrane (M) protein, has a pivotal role both during viral infection cycle and host interferon antagonism. This is a highly conserved viral protein, thus an interesting and suitable target for drug discovery. In this paper, we explain the structural nature of M protein homodimer. To do so, we developed and applied a detailed and robust in silico workflow to predict M protein dimeric structure, membrane orientation, and interface characterization. Single Nucleotide Polymorphisms (SNPs) in M protein were retrieved from over 1.2 M SARS-CoV-2 genomes and proteins from the Global Initiative on Sharing All Influenza Data (GISAID) database, 91 of which were located at the predicted dimer interface. Among those, we identified SNPs in Variants of Concern (VOC) and Variants of Interest (VOI). Binding free energy differences were evaluated for dimer interfacial SNPs to infer mutant protein stabilities. A few high-prevalent mutated residues were found to be especially relevant in VOC and VOI. This realization may be a game-changer to structure-driven formulation of new therapeutics for SARS-CoV-2.

Artificial intelligence based methods for hot spot prediction

Proteins interact through their interfaces to fulfill essential functions in the cell. They bind to their partners in a highly specific manner and form complexes that have a profound effect on understanding the biological pathways they are involved in. Any abnormal interactions may cause diseases. Therefore, the identification of small molecules which modulate protein interactions through their interfaces has high therapeutic potential. However, discovering such molecules is challenging. Most protein-protein binding affinity is attributed to a small set of amino acids found in protein interfaces known as hot spots. Recent studies demonstrate that drug-like small molecules specifically may bind to hot spots. Therefore, hot spot prediction is crucial. As experimental data accumulates, artificial intelligence begins to be used for computational hot spot prediction. First, we review machine learning and deep learning for computational hot spot prediction and then explain the significance of hot spots toward drug design.

In Silico End-to-End Protein-Ligand Interaction Characterization Pipeline: The Case of SARS-CoV-2

SARS-CoV-2 triggered a worldwide pandemic disease, COVID-19, for which an effective treatment has not yet been settled. Among the most promising targets to fight this disease is SARS-CoV-2 main protease (Mpro), which has been extensively studied in the last few months. There is an urgency for developing effective computational protocols that can help us tackle these key viral proteins. Hence, we have put together a robust and thorough pipeline of in silico protein-ligand characterization methods to address one of the biggest biological problems currently plaguing our world. These methodologies were used to characterize the interaction of SARS-CoV-2 Mpro with an α-ketoamide inhibitor and include details on how to upload, visualize, and manage the three-dimensional structure of the complex and acquire high-quality figures for scientific publications using PyMOL (Protocol 1); perform homology modeling with MODELLER (Protocol 2); perform protein-ligand docking calculations using HADDOCK (Protocol 3); run a virtual screening protocol of a small compound database of SARS-CoV-2 candidate inhibitors with AutoDock 4 and AutoDock Vina (Protocol 4); and, finally, sample the conformational space at the atomic level between SARS-CoV-2 Mpro and the α-ketoamide inhibitor with Molecular Dynamics simulations using GROMACS (Protocol 5). Guidelines for careful data analysis and interpretation are also provided for each Protocol.

Protein-protein interactions at a glance: Protocols for the visualization of biomolecular interactions

Protein-protein interactions (PPIs) play a key role in many biological processes and are intriguing targets for drug discovery campaigns. Advancements in experimental and computational techniques are leading to a growth of data accessibility, and, with it, an increased need for the analysis of PPIs. In this respect, visualization tools are essential instruments to represent and analyze biomolecular interactions. In this chapter, we reviewed some of the available tools, highlighting their features, and describing their functions with practical information on their usage.