Exploiting pretrained biochemical language models for targeted drug design.

MOTIVATION: The development of novel compounds targeting proteins of interest is one of the most important tasks in the pharmaceutical industry. Deep generative models have been applied to targeted molecular design and have shown promising results. Recently, target-specific molecule generation has been viewed as a translation between the protein language and the chemical language. However, such a model is limited by the availability of interacting protein-ligand pairs. On the other hand, large amounts of unlabelled protein sequences and chemical compounds are available and have been used to train language models that learn useful representations. In this study, we propose exploiting pretrained biochemical language models to initialize (i.e. warm start) targeted molecule generation models. We investigate two warm start strategies: (i) a one-stage strategy where the initialized model is trained on targeted molecule generation and (ii) a two-stage strategy containing a pre-finetuning on molecular generation followed by target-specific training. We also compare two decoding strategies to generate compounds: beam search and sampling. RESULTS: The results show that the warm-started models perform better than a baseline model trained from scratch. The two proposed warm-start strategies achieve similar results to each other with respect to widely used metrics from benchmarks. However, docking evaluation of the generated compounds for a number of novel proteins suggests that the one-stage strategy generalizes better than the two-stage strategy. Additionally, we observe that beam search outperforms sampling in both docking evaluation and benchmark metrics for assessing compound quality. AVAILABILITY AND IMPLEMENTATION: The source code is available at https://github.com/boun-tabi/biochemical-lms-for-drug-design and the materials (i.e., data, models, and outputs) are archived in Zenodo at https://doi.org/10.5281/zenodo.6832145. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Carvacrol Enhances the Antimicrobial Potency of Berberine in Bacillus subtilis.

The essential oil carvacrol from oregano displays a wide range of biological activities among which is found the inhibition of efflux pumps. Thus, using carvacrol, the current work undertook the effort to potentiate the antimicrobial activity of berberine, a natural product with limited antimicrobial efficacy due to its efflux. Following the selection of concentrations for the combinatorial treatments, guided by checkerboard microtiter plate assay and growth experiments, ethidium bromide accumulation assay was used to find that 25 µg mL-1 carvacrol displayed a weak efflux pump inhibitor character in Bacillus subtilis. Scanning electron microscopy images and cellular material leakage assays showed that carvacrol at this concentration neither altered the morphology nor the permeability of the membrane alone but when combined with 75 µg mL-1 berberine. Among the efflux pumps of different families found in B. subtilis, except for BmrA and Mdr, the increase in the expressional changes was striking, with Blt displaying ~ 4500-fold increase in expression under the combination treatment. Overall, the findings demonstrated that carvacrol potentiated the effect of berberine; however, not only multiple pumps but also different targets may be responsible for the observed activity.

Superior inhibition of influenza virus hemagglutinin-mediated fusion by indole-substituted spirothiazolidinones.

The influenza virus hemagglutinin (HA) mediates membrane fusion after viral entry by endocytosis. The fusion process requires drastic low pH-induced HA refolding and is prevented by arbidol and tert-butylhydroquinone (TBHQ). We here report a class of superior inhibitors with indole-substituted spirothiazolidinone structure. The most active analogue 5f has an EC50 value against influenza A/H3N2 virus of 1 nM and selectivity index of almost 2000. Resistance data and in silico modeling indicate that 5f combines optimized fitting in the TBHQ/arbidol HA binding pocket with a capability for endosomal accumulation. Both criteria appear relevant to achieve superior inhibitors of HA-mediated fusion.

A novel methodology on distributed representations of proteins using their interacting ligands.

Motivation: The effective representation of proteins is a crucial task that directly affects the performance of many bioinformatics problems. Related proteins usually bind to similar ligands. Chemical characteristics of ligands are known to capture the functional and mechanistic properties of proteins suggesting that a ligand-based approach can be utilized in protein representation. In this study, we propose SMILESVec, a Simplified molecular input line entry system (SMILES)-based method to represent ligands and a novel method to compute similarity of proteins by describing them based on their ligands. The proteins are defined utilizing the word-embeddings of the SMILES strings of their ligands. The performance of the proposed protein description method is evaluated in protein clustering task using TransClust and MCL algorithms. Two other protein representation methods that utilize protein sequence, Basic local alignment tool and ProtVec, and two compound fingerprint-based protein representation methods are compared. Results: We showed that ligand-based protein representation, which uses only SMILES strings of the ligands that proteins bind to, performs as well as protein sequence-based representation methods in protein clustering. The results suggest that ligand-based protein description can be an alternative to the traditional sequence or structure-based representation of proteins and this novel approach can be applied to different bioinformatics problems such as prediction of new protein-ligand interactions and protein function annotation. Availability and implementation: https://github.com/hkmztrk/SMILESVecProteinRepresentation. Supplementary information: Supplementary data are available at Bioinformatics online.

DeepDTA: deep drug-target binding affinity prediction.

Motivation: The identification of novel drug-target (DT) interactions is a substantial part of the drug discovery process. Most of the computational methods that have been proposed to predict DT interactions have focused on binary classification, where the goal is to determine whether a DT pair interacts or not. However, protein-ligand interactions assume a continuum of binding strength values, also called binding affinity and predicting this value still remains a challenge. The increase in the affinity data available in DT knowledge-bases allows the use of advanced learning techniques such as deep learning architectures in the prediction of binding affinities. In this study, we propose a deep-learning based model that uses only sequence information of both targets and drugs to predict DT interaction binding affinities. The few studies that focus on DT binding affinity prediction use either 3D structures of protein-ligand complexes or 2D features of compounds. One novel approach used in this work is the modeling of protein sequences and compound 1D representations with convolutional neural networks (CNNs). Results: The results show that the proposed deep learning based model that uses the 1D representations of targets and drugs is an effective approach for drug target binding affinity prediction. The model in which high-level representations of a drug and a target are constructed via CNNs achieved the best Concordance Index (CI) performance in one of our larger benchmark datasets, outperforming the KronRLS algorithm and SimBoost, a state-of-the-art method for DT binding affinity prediction. Availability and implementation: https://github.com/hkmztrk/DeepDTA. Supplementary information: Supplementary data are available at Bioinformatics online.

pVEC hydrophobic N-terminus is critical for antibacterial activity.

Cell-penetrating peptides (CPPs) are commonly defined by their shared ability to be internalized into eukaryotic cells, without inducing permanent membrane damage, and to improve cargo delivery. Many CPPs also possess antimicrobial action strong enough to selectively lyse microbes in infected mammalian cultures. pVEC, a CPP derived from cadherin, is able to translocate into mammalian cells, and it is also antimicrobial. Structure-activity relationship and sequence alignment studies have suggested that the hydrophobic N-terminus (LLIIL) of pVEC is essential for this peptide's uptake into eukaryotic cells. In this study, our aim was to examine the contribution of these residues to the antimicrobial action and the translocation mechanism of pVEC. We performed antimicrobial activity and microscopy experiments with pVEC and with del5 pVEC (N-terminal truncated variant of pVEC) and showed that pVEC loses its antimicrobial effect upon deletion of the LLIIL residues, even though both peptides induce membrane permeability. We also calculated the free energy of the transport process using steered molecular dynamic simulations and replica exchange umbrella sampling simulations to compare the difference in uptake mechanism of the 2 peptides in atomistic detail. Despite the difference in experimentally observed antimicrobial activity, the simulations on the 2 peptides showed similar characteristics and the energetic cost of translocation of pVEC was higher than that of del5 pVEC, suggesting that pVEC uptake mechanism cannot be explained by simple passive transport. Our results suggest that LLIIL residues are key contributors to pVEC antibacterial activity because of irreversible membrane disruption.

The effect of a beta-lactamase inhibitor peptide on bacterial membrane structure and integrity: a comparative study.

Co-administration of beta-lactam antibiotics and beta-lactamase inhibitors has been a favored treatment strategy against beta-lactamase-mediated bacterial antibiotic resistance, but the emergence of beta-lactamases resistant to current inhibitors necessitates the discovery of novel non-beta-lactam inhibitors. Peptides derived from the Ala46-Tyr51 region of the beta-lactamase inhibitor protein are considered as potent inhibitors of beta-lactamase; unfortunately, peptide delivery into the cell limits their potential. The properties of cell-penetrating peptides could guide the design of beta-lactamase inhibitory peptides. Here, our goal is to modify the peptide with the sequence RRGHYY that possesses beta-lactamase inhibitory activity under in vitro conditions. Inspired by the work on the cell-penetrating peptide pVEC, our approach involved the addition of the N-terminal hydrophobic residues, LLIIL, from pVEC to the inhibitor peptide to build a chimera. These residues have been reported to be critical in the uptake of pVEC. We tested the potential of RRGHYY and its chimeric derivative as a beta-lactamase inhibitory peptide on Escherichia coli cells and compared the results with the action of the antimicrobial peptide melittin, the beta-lactam antibiotic ampicillin, and the beta-lactamase inhibitor potassium clavulanate to get mechanistic details on their action. Our results show that the addition of LLIIL to the N-terminus of the beta-lactamase inhibitory peptide RRGHYY increases its membrane permeabilizing potential. Interestingly, the addition of this short stretch of hydrophobic residues also modified the inhibitory peptide such that it acquired antimicrobial property. We propose that addition of the hydrophobic LLIIL residues to the peptide N-terminus offers a promising strategy to design novel antimicrobial peptides in the battle against antibiotic resistance. Copyright © 2017 European Peptide Society and John Wiley & Sons, Ltd.

A comparative study of SMILES-based compound similarity functions for drug-target interaction prediction.

BACKGROUND: Molecular structures can be represented as strings of special characters using SMILES. Since each molecule is represented as a string, the similarity between compounds can be computed using SMILES-based string similarity functions. Most previous studies on drug-target interaction prediction use 2D-based compound similarity kernels such as SIMCOMP. To the best of our knowledge, using SMILES-based similarity functions, which are computationally more efficient than the 2D-based kernels, has not been investigated for this task before. RESULTS: In this study, we adapt and evaluate various SMILES-based similarity methods for drug-target interaction prediction. In addition, inspired by the vector space model of Information Retrieval we propose cosine similarity based SMILES kernels that make use of the Term Frequency (TF) and Term Frequency-Inverse Document Frequency (TF-IDF) weighting approaches. We also investigate generating composite kernels by combining our best SMILES-based similarity functions with the SIMCOMP kernel. With this study, we provided a comparison of 13 different ligand similarity functions, each of which utilizes the SMILES string of molecule representation. Additionally, TF and TF-IDF based cosine similarity kernels are proposed. CONCLUSION: The more efficient SMILES-based similarity functions performed similarly to the more complex 2D-based SIMCOMP kernel in terms of AUC-ROC scores. The TF-IDF based cosine similarity obtained a better AUC-PR score than the SIMCOMP kernel on the GPCR benchmark data set. The composite kernel of TF-IDF based cosine similarity and SIMCOMP achieved the best AUC-PR scores for all data sets.

A novel chimeric peptide with antimicrobial activity.

Beta-lactamase-mediated bacterial drug resistance exacerbates the prognosis of infectious diseases, which are sometimes treated with co-administration of beta-lactam type antibiotics and beta-lactamase inhibitors. Antimicrobial peptides are promising broad-spectrum alternatives to conventional antibiotics in this era of evolving bacterial resistance. Peptides based on the Ala46-Tyr51 beta-hairpin loop of beta-lactamase inhibitory protein (BLIP) have been previously shown to inhibit beta-lactamase. Here, our goal was to modify this peptide for improved beta-lactamase inhibition and cellular uptake. Motivated by the cell-penetrating pVEC sequence, which includes a hydrophobic stretch at its N-terminus, our approach involved the addition of LLIIL residues to the inhibitory peptide N-terminus to facilitate uptake. Activity measurements of the peptide based on the 45-53 loop of BLIP for enhanced inhibition verified that the peptide was a competitive beta-lactamase inhibitor with a K(i) value of 58 µM. Incubation of beta-lactam-resistant cells with peptide decreased the number of viable cells, while it had no effect on beta-lactamase-free cells, indicating that this peptide had antimicrobial activity via beta-lactamase inhibition. To elucidate the molecular mechanism by which this peptide moves across the membrane, steered molecular dynamics simulations were carried out. We propose that addition of hydrophobic residues to the N-terminus of the peptide affords a promising strategy in the design of novel antimicrobial peptides not only against beta-lactamase but also for other intracellular targets.

Exploring data-driven chemical SMILES tokenization approaches to identify key protein-ligand binding moieties.

Machine learning models have found numerous successful applications in computational drug discovery. A large body of these models represents molecules as sequences since molecular sequences are easily available, simple, and informative. The sequence-based models often segment molecular sequences into pieces called chemical words, analogous to the words that make up sentences in human languages, and then apply advanced natural language processing techniques for tasks such as de novo drug design, property prediction, and binding affinity prediction. However, the chemical characteristics and significance of these building blocks, chemical words, remain unexplored. To address this gap, we employ data-driven SMILES tokenization techniques such as Byte Pair Encoding, WordPiece, and Unigram to identify chemical words and compare the resulting vocabularies. To understand the chemical significance of these words, we build a language-inspired pipeline that treats high affinity ligands of protein targets as documents and selects key chemical words making up those ligands based on tf-idf weighting. The experiments on multiple protein-ligand affinity datasets show that despite differences in words, lengths, and validity among the vocabularies generated by different subword tokenization algorithms, the identified key chemical words exhibit similarity. Further, we conduct case studies on a number of target to analyze the impact of key chemical words on binding. We find that these key chemical words are specific to protein targets and correspond to known pharmacophores and functional groups. Our approach elucidates chemical properties of the words identified by machine learning models and can be used in drug discovery studies to determine significant chemical moieties.

Machine-learning-based adverse drug event prediction from observational health data: A review.

Adverse drug events (ADEs) are responsible for a significant number of hospital admissions and fatalities. Machine learning models have been developed to assess the individual patient risk of having an ADE. In this article, we have reviewed studies addressing the prediction of ADEs in observational health data with machine learning. The field of individualised ADE prediction is rapidly emerging through the increasing availability of additional data modalities (e.g., genetic data, screening data, wearables data) and advanced deep learning models such as transformers. Consequently, personalised adverse drug event predictions are becoming more feasible and tangible.

A Computational Software for Training Robust Drug-Target Affinity Prediction Models: pydebiaseddta.

Robust generalization of drug-target affinity (DTA) prediction models is a notoriously difficult problem in computational drug discovery. In this article, we present pydebiaseddta: a computational software for improving the generalizability of DTA prediction models to novel ligands and/or proteins. pydebiaseddta serves as the practical implementation of the DebiasedDTA training framework, which advocates modifying the training distribution to mitigate the effect of spurious correlations in the training data set that leads to substantially degraded performance for novel ligands and proteins. Written in Python programming language, pydebiaseddta combines a user-friendly streamlined interface with a feature-rich and highly modifiable architecture. With this article we introduce our software, showcase its main functionalities, and describe practical ways for new users to engage with it.

A Framework for Improving the Generalizability of Drug-Target Affinity Prediction Models.

Statistical models that accurately predict the binding affinity of an input ligand-protein pair can greatly accelerate drug discovery. Such models are trained on available ligand-protein interaction data sets, which may contain biases that lead the predictor models to learn data set-specific, spurious patterns instead of generalizable relationships. This leads the prediction performances of these models to drop dramatically for previously unseen biomolecules. Various approaches that aim to improve model generalizability either have limited applicability or introduce the risk of degrading overall prediction performance. In this article, we present DebiasedDTA, a novel training framework for drug-target affinity (DTA) prediction models that addresses data set biases to improve the generalizability of such models. DebiasedDTA relies on reweighting the training samples to achieve robust generalization, and is thus applicable to most DTA prediction models. Extensive experiments with different biomolecule representations, model architectures, and data sets demonstrate that DebiasedDTA achieves improved generalizability in predicting drug-target affinities.

Automatic Extraction of Medication Mentions from Tweets-Overview of the BioCreative VII Shared Task 3 Competition.

This study presents the outcomes of the shared task competition BioCreative VII (Task 3) focusing on the extraction of medication names from a Twitter user's publicly available tweets (the user's 'timeline'). In general, detecting health-related tweets is notoriously challenging for natural language processing tools. The main challenge, aside from the informality of the language used, is that people tweet about any and all topics, and most of their tweets are not related to health. Thus, finding those tweets in a user's timeline that mention specific health-related concepts such as medications requires addressing extreme imbalance. Task 3 called for detecting tweets in a user's timeline that mentions a medication name and, for each detected mention, extracting its span. The organizers made available a corpus consisting of 182 049 tweets publicly posted by 212 Twitter users with all medication mentions manually annotated. The corpus exhibits the natural distribution of positive tweets, with only 442 tweets (0.2%) mentioning a medication. This task was an opportunity for participants to evaluate methods that are robust to class imbalance beyond the simple lexical match. A total of 65 teams registered, and 16 teams submitted a system run. This study summarizes the corpus created by the organizers and the approaches taken by the participating teams for this challenge. The corpus is freely available at https://biocreative.bioinformatics.udel.edu/tasks/biocreative-vii/track-3/. The methods and the results of the competing systems are analyzed with a focus on the approaches taken for learning from class-imbalanced data.

Labels in a haystack: Approaches beyond supervised learning in biomedical applications.

Recent advances in biomedical machine learning demonstrate great potential for data-driven techniques in health care and biomedical research. However, this potential has thus far been hampered by both the scarcity of annotated data in the biomedical domain and the diversity of the domain's subfields. While unsupervised learning is capable of finding unknown patterns in the data by design, supervised learning requires human annotation to achieve the desired performance through training. With the latter performing vastly better than the former, the need for annotated datasets is high, but they are costly and laborious to obtain. This review explores a family of approaches existing between the supervised and the unsupervised problem setting. The goal of these algorithms is to make more efficient use of the available labeled data. The advantages and limitations of each approach are addressed and perspectives are provided.

ChemBoost: A Chemical Language Based Approach for Protein - Ligand Binding Affinity Prediction.

Identification of high affinity drug-target interactions is a major research question in drug discovery. Proteins are generally represented by their structures or sequences. However, structures are available only for a small subset of biomolecules and sequence similarity is not always correlated with functional similarity. We propose ChemBoost, a chemical language based approach for affinity prediction using SMILES syntax. We hypothesize that SMILES is a codified language and ligands are documents composed of chemical words. These documents can be used to learn chemical word vectors that represent words in similar contexts with similar vectors. In ChemBoost, the ligands are represented via chemical word embeddings, while the proteins are represented through sequence-based features and/or chemical words of their ligands. Our aim is to process the patterns in SMILES as a language to predict protein-ligand affinity, even when we cannot infer the function from the sequence. We used eXtreme Gradient Boosting to predict protein-ligand affinities in KIBA and BindingDB data sets. ChemBoost was able to predict drug-target binding affinity as well as or better than state-of-the-art machine learning systems. When powered with ligand-centric representations, ChemBoost was more robust to the changes in protein sequence similarity and successfully captured the interactions between a protein and a ligand, even if the protein has low sequence similarity to the known targets of the ligand.

Exploring chemical space using natural language processing methodologies for drug discovery.

Text-based representations of chemicals and proteins can be thought of as unstructured languages codified by humans to describe domain-specific knowledge. Advances in natural language processing (NLP) methodologies in the processing of spoken languages accelerated the application of NLP to elucidate hidden knowledge in textual representations of these biochemical entities and then use it to construct models to predict molecular properties or to design novel molecules. This review outlines the impact made by these advances on drug discovery and aims to further the dialogue between medicinal chemists and computer scientists.

Synthesis and biological evaluation of indole-2-carbohydrazides and thiazolidinyl-indole-2-carboxamides as potent tubulin polymerization inhibitors.

A new series of N'-(substituted phenyl)-5-chloro/iodo-3-phenyl-1H-indole-2-carbohydrazide (5, 6) and N-[2-(substituted phenyl)-4-oxo-1,3-thiazolidin-3-yl]-5-iodo/chloro-3-phenyl-1H-indole-2-carboxamide (7, 8) derivatives were synthesized and evaluated for their anticancer properties. Compounds 5a and 6b, selected as prototypes by the National Cancer Institute for screening against the full panel of 60 human tumor cell lines at a minimum of five concentrations at 10-fold dilutions, demonstrated remarkable antiproliferative activity against leukemia, non-small cell lung cancer, colon cancer, central nervous system (CNS) cancer, melanoma, ovarian cancer, renal cancer, and breast cancer (MCF-7) cell lines with GI50 values < 0.4 µM. A subset of the compounds was then tested for their potential to inhibit tubulin polymerization. Compounds 6f and 6g showed significant cytotoxicity at the nM level on MCF-7 cells and exhibited significant inhibitory activity on tubulin assembly and colchicine binding at about the same level as combretastatin A-4. Finally, docking calculations were performed to identify the binding mode of these compounds. Group 5 and 6 compounds interacted with the colchicine binding site through hydrophobic interactions similar to those of colchicine. These compounds with antiproliferative activity at high nanomolar concentration can serve as scaffolds for the design of novel microtubule targeting agents.

In Silico Identification of Novel Orthosteric Inhibitors of Sphingosine Kinase 1 (SK1).

BACKGROUND: Sphingosine kinase 1 (SK1) overexpression and elevated sphingosine-1-phosphate (S1P) levels have been correlated with many disease states from cancer to inflammatory diseases to diabetes. Even though SK1 inhibitors are of consideberable interest as effective chemotherapeutic agents, poor potency, lack of selectivity and poor pharmacokinetic properties have been major problems in the first generation SK1 inhibitors. OBJECTIVE: There is an urgent need for the discovery of novel in vivo, stable selective SK1 inhibitors with improved potency. The primary object of this study was to identify potential novel leads for orthosteric inhibition of SK1. METHODS: We propose a series of compounds from different chemotypes as potential selective SK1 inhibitors via virtual screening of the ZINC database using ligand-based and structure-based pharmacophore models, molecular docking, substructure search, selectivity calculations. Molecular dynamics (MD) simulations revealed key insights into the binding mode and the stability of the SK1-ligand complex. RESULTS: Ten ligands were proposed as potential SK1 inhibitors based on the high induced fit docking scores, BEI, LLE and %HOA. Ligands 2, 3, 5 and 9 were found to be selective toward SK1 with favorable binding free energy of - 95 ± 5 kcal/mol. MD simulation of ligand 5 showed that the ligand-SK1 complex reached equilibrium with favorable hydrogen bonding and hydrophobic interactions. The four selective compounds have less than 0.24 similarity with previously discovered potent inhibitors. CONCLUSION: The proposed compounds may serve as potential novel leads for orthosteric inhibition of SK1.

Membrane Active Peptides and Their Biophysical Characterization.

In the last 20 years, an increasing number of studies have been reported on membrane active peptides. These peptides exert their biological activity by interacting with the cell membrane, either to disrupt it and lead to cell lysis or to translocate through it to deliver cargos into the cell and reach their target. Membrane active peptides are attractive alternatives to currently used pharmaceuticals and the number of antimicrobial peptides (AMPs) and peptides designed for drug and gene delivery in the drug pipeline is increasing. Here, we focus on two most prominent classes of membrane active peptides; AMPs and cell-penetrating peptides (CPPs). Antimicrobial peptides are a group of membrane active peptides that disrupt the membrane integrity or inhibit the cellular functions of bacteria, virus, and fungi. Cell penetrating peptides are another group of membrane active peptides that mainly function as cargo-carriers even though they may also show antimicrobial activity. Biophysical techniques shed light on peptide⁻membrane interactions at higher resolution due to the advances in optics, image processing, and computational resources. Structural investigation of membrane active peptides in the presence of the membrane provides important clues on the effect of the membrane environment on peptide conformations. Live imaging techniques allow examination of peptide action at a single cell or single molecule level. In addition to these experimental biophysical techniques, molecular dynamics simulations provide clues on the peptide⁻lipid interactions and dynamics of the cell entry process at atomic detail. In this review, we summarize the recent advances in experimental and computational investigation of membrane active peptides with particular emphasis on two amphipathic membrane active peptides, the AMP melittin and the CPP pVEC.