Improving Docking Power for Short Peptides Using Random Forest

Genome sequencing and protein modeling unraveled the 2AP biosynthesis in Bacillus cereus DB25

2-Acetyl-1-pyrroline (2AP) is an important and major flavor aroma compound responsible for the fragrance of basmati rice, cheese, wine, and several other food products. Biosynthesis of 2AP in aromatic rice and a few other plant species is associated with a recessive Betaine aldehyde dehydrogenase 2 (BADH2) gene. However, the literature is scant on the relationship between the functional BADH2 gene and 2AP biosynthesis in prokaryotic systems. Therefore, in the present study, we aimed to explore the functionality of the BADH2 gene for 2AP biosynthesis in 2AP synthesizing rice rhizobacterial isolate Bacillus cereus DB25 isolated from the rhizosphere of basmati rice (Oryza sativa L.). Full-length BcBADH2 sequence was obtained through whole genome sequencing (WGS) and further confirmed through traditional PCR and Sanger sequencing. Then the functionality of the BcBADH2 gene was evaluated in-silico through bioinformatics analysis and protein docking studies and further experimentally validated through enzyme assay. The sequencing and bioinformatics analysis results revealed a full-length 1485 bp BcBADH2 coding sequence without any deletion or premature stop codons. Full-length BcBADH2 was found to encode a fully functional protein of 54.08 kDa with pI of 5.22 and showed the presence of the conserved amino acids responsible for enzyme activity. The docking studies confirmed a good affinity between the protein and its substrate whereas the presence of BcBADH2 enzyme activity confirmed the functionality of BADH2 enzyme in B. cereus DB25. In conclusion, the findings of the present study suggest that B. cereus DB25 is able to synthesize 2AP despite a functional BADH2 gene and there may be a different molecular mechanism responsible for 2AP biosynthesis in bacterial systems, unlike that found in aromatic rice and other eukaryotic plant species.

How do physicochemical properties contribute to inhibitory activity of promising peptides against Zika Virus NS3 protease?

CONTEXT AND RESULTS

Flavivirus diseases' cycles, especially Dengue and Yellow Fever, can be observed all over Brazilian territory, representing a great health concern. Additionally, there are no drugs available in therapy. In this scenario, in silico methodologies were applied to obtain physicochemical properties, as well as to better understand the ligand-biological target interaction mode of 20 previously reported NS2B/NS3 protease inhibitors of Dengue virus. Since catalytic site of flavivirus hold similarities, such as the same catalytic triad (His51, Asp75 e Ser135), the ability of this series of molecules to fit in Zika NS3 domains can be achieved. We performed an exploratory data analysis, using statistical methodologies, such as PCA (Principal Component Analysis) and HCA (Hierarchical Component Analysis), to assist the comprehension of how physicochemical properties impact the interaction observed by the docking studies, as well as to build a correlation between the respective ranked characteristics. Based on these previous studies, peptides were selected for the dynamics simulations, which were useful to better understand the ligand-protein interactions. Information relating to, for instance, energy, ΔG, average number of hydrogen bonds and distance from Ser135 (one of the main amino acids in the catalytic pocket) were discussed. In this sense, peptides 15 (considering ΔG value and Hbond number), 7 (ΔG and energy) and 1, 6, 7 and 15 (the proximity to Ser135 throughout the dynamics simulation) were highlighted as promising. Those interesting results could contribute to future studies regarding Zika virus drug design, since this infection represents a great concern in neglected populations.

METHODS

The models were constructed in the ChemDraw software. The ligand parametrization was performed in the CHEM3D 17.0, UCSF Chimera. Docking simulations were carried out in the GOLD software, after the redocking validation. We used ASP as the function score. Additionally, for dynamics simulations we applied GROMACS software, exploring, mainly, free binding energy calculations. Exploratory analysis was carried out in Minitab 17.3.1 statistical software. Prior to the exploratory analysis, data of quantum chemical properties of the peptides were collected in Microsoft Excel spreadsheet and organized to obtain Hierarchical Cluster Analysis (HCA) and Principal Component Analysis (PCA).

Identification of defactinib derivatives targeting focal adhesion kinase using ensemble docking, molecular dynamics simulations and binding free energy calculations

Focal adhesion kinase (FAK) belongs to the nonreceptor tyrosine kinases, which selectively phosphorylate tyrosine residues on substrate proteins. FAK is associated with bladder, esophageal, gastric, neck, breast, ovarian and lung cancers. Thus, FAK has been considered as a potential target for tumor treatment. Currently, there are six adenosine triphosphate (ATP)-competitive FAK inhibitors tested in clinical trials but no approved inhibitors targeting FAK. Defactinib (VS-6063) is a second-generation FAK inhibitor with an IC50 of 0.6 nM. The binding model of VS-6063 with FAK may provide a reference model for developing new antitumor FAK-targeting drugs. In this study, the VS-6063/FAK binding model was constructed using ensemble docking and molecular dynamics simulations. Furthermore, the molecular mechanics/generalized Born (GB) surface area (MM/GBSA) method was employed to estimate the binding free energy between VS-6063 and FAK. The key residues involved in VS-6063/FAK binding were also determined using per-residue energy decomposition analysis. Based on the binding model, VS-6063 could be separated into seven regions to enhance its binding affinity with FAK. Meanwhile, 60 novel defactinib-based compounds were designed and verified using ensemble docking. Overall, the present study improves our understanding of the binding mechanism of human FAK with VS-6063 and provides new insights into future drug designs targeting FAK.Communicated by Ramaswamy H. Sarma.

Machine Learning Methods for Small Data Challenges in Molecular Science

Small data are often used in scientific and engineering research due to the presence of various constraints, such as time, cost, ethics, privacy, security, and technical limitations in data acquisition. However, big data have been the focus for the past decade, small data and their challenges have received little attention, even though they are technically more severe in machine learning (ML) and deep learning (DL) studies. Overall, the small data challenge is often compounded by issues, such as data diversity, imputation, noise, imbalance, and high-dimensionality. Fortunately, the current big data era is characterized by technological breakthroughs in ML, DL, and artificial intelligence (AI), which enable data-driven scientific discovery, and many advanced ML and DL technologies developed for big data have inadvertently provided solutions for small data problems. As a result, significant progress has been made in ML and DL for small data challenges in the past decade. In this review, we summarize and analyze several emerging potential solutions to small data challenges in molecular science, including chemical and biological sciences. We review both basic machine learning algorithms, such as linear regression, logistic regression (LR), k-nearest neighbor (KNN), support vector machine (SVM), kernel learning (KL), random forest (RF), and gradient boosting trees (GBT), and more advanced techniques, including artificial neural network (ANN), convolutional neural network (CNN), U-Net, graph neural network (GNN), Generative Adversarial Network (GAN), long short-term memory (LSTM), autoencoder, transformer, transfer learning, active learning, graph-based semi-supervised learning, combining deep learning with traditional machine learning, and physical model-based data augmentation. We also briefly discuss the latest advances in these methods. Finally, we conclude the survey with a discussion of promising trends in small data challenges in molecular science.

Dock-able linear and homodetic di, tri, tetra and pentapeptide library from canonical amino acids: SARS-CoV-2 Mpro as a case study

Peptide-based therapeutics are increasingly pushing to the forefront of biomedicine with their promise of high specificity and low toxicity. Although noncanonical residues can always be used, employing only the natural 20 residues restricts the chemical space to a finite dimension allowing for comprehensive in silico screening. Towards this goal, the dataset comprising all possible di-, tri-, and tetra-peptide combinations of the canonical residues has been previously reported. However, with increasing computational power, the comprehensive set of pentapeptides is now also feasible for screening as the comprehensive set of cyclic peptides comprising four or five residues. Here, we provide both the complete and prefiltered libraries of all di-, tri-, tetra-, and penta-peptide sequences from 20 canonical amino acids and their homodetic (N-to-C-terminal) cyclic homologues. The FASTA, simplified molecular-input line-entry system (SMILES), and structure-data file (SDF)-three dimension (3D) libraries can be readily used for screening against protein targets. We also provide a simple method and tool for conducting identity-based filtering. Access to this dataset will accelerate small peptide screening workflows and encourage their use in drug discovery campaigns. As a case study, the developed library was screened against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) main protease to identify potential small peptide inhibitors.

Recent PELE Developments and Applications in Drug Discovery Campaigns

Computer simulation techniques are gaining a central role in molecular pharmacology. Due to several factors, including the significant improvements of traditional molecular modelling, the irruption of machine learning methods, the massive data generation, or the unlimited computational resources through cloud computing, the future of pharmacology seems to go hand in hand with in silico predictions. In this review, we summarize our recent efforts in such a direction, centered on the unconventional Monte Carlo PELE software and on its coupling with machine learning techniques. We also provide new data on combining two recent new techniques, aquaPELE capable of exhaustive water sampling and fragPELE, for fragment growing.

Application of CT-Based Radiomics in Discriminating Pancreatic Cystadenomas From Pancreatic Neuroendocrine Tumors Using Machine Learning Methods

Objectives

The purpose of this study aimed at investigating the reliability of radiomics features extracted from contrast-enhanced CT in differentiating pancreatic cystadenomas from pancreatic neuroendocrine tumors (PNETs) using machine-learning methods.

Methods

In this study, a total number of 120 patients, including 66 pancreatic cystadenomas patients and 54 PNETs patients were enrolled. Forty-eight radiomic features were extracted from contrast-enhanced CT images using LIFEx software. Five feature selection methods were adopted to determine the appropriate features for classifiers. Then, nine machine learning classifiers were employed to build predictive models. The performance of the forty-five models was evaluated with area under the curve (AUC), accuracy, sensitivity, specificity, and F1 score in the testing group.

Results

The predictive models exhibited reliable ability of differentiating pancreatic cystadenomas from PNETs when combined with suitable selection methods. A combination of DC as the selection method and RF as the classifier, as well as Xgboost+RF, demonstrated the best discriminative ability, with the highest AUC of 0.997 in the testing group.

Conclusions

Radiomics-based machine learning methods might be a noninvasive tool to assist in differentiating pancreatic cystadenomas and PNETs.