Biomolecular modeling thrives in the age of technology

Ab initio characterization of protein molecular dynamics with AI2BMD

Biomolecular dynamics simulation is a fundamental technology for life sciences research, and its usefulness depends on its accuracy and efficiency1-3. Classical molecular dynamics simulation is fast but lacks chemical accuracy4,5. Quantum chemistry methods such as density functional theory can reach chemical accuracy but cannot scale to support large biomolecules6. Here we introduce an artificial intelligence-based ab initio biomolecular dynamics system (AI2BMD) that can efficiently simulate full-atom large biomolecules with ab initio accuracy. AI2BMD uses a protein fragmentation scheme and a machine learning force field7 to achieve generalizable ab initio accuracy for energy and force calculations for various proteins comprising more than 10,000 atoms. Compared to density functional theory, it reduces the computational time by several orders of magnitude. With several hundred nanoseconds of dynamics simulations, AI2BMD demonstrated its ability to efficiently explore the conformational space of peptides and proteins, deriving accurate 3J couplings that match nuclear magnetic resonance experiments, and showing protein folding and unfolding processes. Furthermore, AI2BMD enables precise free-energy calculations for protein folding, and the estimated thermodynamic properties are well aligned with experiments. AI2BMD could potentially complement wet-lab experiments, detect the dynamic processes of bioactivities and enable biomedical research that is impossible to conduct at present.

Navigating the landscape of enzyme design: from molecular simulations to machine learning

Global environmental issues and sustainable development call for new technologies for fine chemical synthesis and waste valorization. Biocatalysis has attracted great attention as the alternative to the traditional organic synthesis. However, it is challenging to navigate the vast sequence space to identify those proteins with admirable biocatalytic functions. The recent development of deep-learning based structure prediction methods such as AlphaFold2 reinforced by different computational simulations or multiscale calculations has largely expanded the 3D structure databases and enabled structure-based design. While structure-based approaches shed light on site-specific enzyme engineering, they are not suitable for large-scale screening of potential biocatalysts. Effective utilization of big data using machine learning techniques opens up a new era for accelerated predictions. Here, we review the approaches and applications of structure-based and machine-learning guided enzyme design. We also provide our view on the challenges and perspectives on effectively employing enzyme design approaches integrating traditional molecular simulations and machine learning, and the importance of database construction and algorithm development in attaining predictive ML models to explore the sequence fitness landscape for the design of admirable biocatalysts.

Advancing biomolecular simulation through exascale HPC, AI and quantum computing

Biomolecular simulation can act as both a digital microscope and a crystal ball; offering the potential for a deeper understanding of experimental observations whilst also presenting a forward-looking avenue for the in silico design and evaluation of hitherto unsynthesized compounds. Indeed, as the intricacy of our scientific inquiries has grown, so too has the computational prowess we seek to deploy in our pursuit of answers. As we enter the Exascale era, this mini-review surveys the computational landscape from both the point of view of the development of new and ever more powerful systems, and the simulations that are run on them. Moreover, as we stand on the cusp of a transformative phase in computational biology, this article offers a contemplative glance into the future, speculating on the profound implications of artificial intelligence and quantum computing for large-scale biomolecular simulations.

Accurate and Efficient Conformer Sampling of Cyclic Drug-Like Molecules with Inverse Kinematics

Identification of all of the influential conformers of biomolecules is a crucial step in many tasks of computational biochemistry. Specifically, molecular docking, a key component of in silico drug development, requires a comprehensive set of conformations for potential candidates in order to generate the optimal ligand-receptor poses and, ultimately, find the best drug candidates. However, the presence of flexible cycles in a molecule complicates the initial search for conformers since exhaustive sampling algorithms via torsional random and systematic searches become very inefficient. The devised inverse-kinematics-based Monte Carlo with refinement (MCR) algorithm identifies independently rotatable dihedral angles in (poly)cyclic molecules and uses them to perform global conformational sampling, outperforming popular alternatives (MacroModel, CREST, and RDKit) in terms of speed and diversity of the resulting conformer ensembles. Moreover, MCR quickly and accurately recovers naturally occurring macrocycle conformations for most of the considered molecules.

Overview of AlphaFold2 and breakthroughs in overcoming its limitations

Predicting three-dimensional (3D) protein structures has been challenging for decades. The emergence of AlphaFold2 (AF2), a deep learning-based machine learning method developed by DeepMind, became a game changer in the protein folding community. AF2 can predict a protein's three-dimensional structure with high confidence based on its amino acid sequence. Accurate prediction of protein structures can dramatically accelerate our understanding of biological mechanisms and provide a solid foundation for reliable drug design. Although AF2 breaks through the barriers in predicting protein structures, many rooms remain to be further studied. This review provides a brief historical overview of the development of protein structure prediction, covering template-based, template-free, and machine learning-based methods. In addition to reviewing the potential benefits (Pros) and considerations (Cons) of using AF2, this review summarizes the diverse applications, including protein structure predictions, dynamic changes, point mutation, integration of language model and experimental data, protein complex, and protein-peptide interaction. It underscores recent advancements in efficiency, reliability, and broad application of AF2. This comprehensive review offers valuable insights into the applications of AF2 and AF2-inspired AI methods in structural biology and its potential for clinically significant drug target discovery.

Molecular choreography: Unveiling the dynamic landscape of type IIA DNA topoisomerases before T-segment passage through all-atom simulations

Type IIA DNA topoisomerases are molecular nanomachines responsible for controlling topological states of DNA molecules. Here, we explore the dynamic landscape of yeast topoisomerase IIA during key stages of its catalytic cycle, focusing in particular on the events preceding the passage of the T-segment. To this end, we generated six configurations of fully catalytic yeast topo IIA, strategically inserted a T-segment into the N-gate in relevant configurations, and performed all-atom simulations. The essential motion of topo IIA protein dimer was characterized by rotational gyrating-like movement together with sliding motion within the DNA-gate. Both appear to be inherent properties of the enzyme and an inbuilt feature that allows passage of the T-segment through the cleaved G-segment. Coupled dynamics of the N-gate and DNA-gate residues may be particularly important for controlled and smooth passage of the T-segment and consequently the prevention of DNA double-strand breaks. QTK loop residue Lys367, which interacts with ATP and ADP molecules, is involved in regulating the size and stability of the N-gate. The unveiled features of the simulated configurations provide insights into the catalytic cycle of type IIA topoisomerases and elucidate the molecular choreography governing their ability to modulate the topological states of DNA topology.

ELViM: Exploring Biomolecular Energy Landscapes through Multidimensional Visualization

Molecular dynamics (MD) simulations provide a powerful means of exploring the dynamic behavior of biomolecular systems at the atomic level. However, analyzing the vast data sets generated by MD simulations poses significant challenges. This article discusses the energy landscape visualization method (ELViM), a multidimensional reduction technique inspired by the energy landscape theory. ELViM transcends one-dimensional representations, offering a comprehensive analysis of the effective conformational phase space without the need for predefined reaction coordinates. We apply the ELViM to study the folding landscape of the antimicrobial peptide Polybia-MP1, showcasing its versatility in capturing complex biomolecular dynamics. Using dissimilarity matrices and a force-scheme approach, the ELViM provides intuitive visualizations, revealing structural correlations and local conformational signatures. The method is demonstrated to be adaptable, robust, and applicable to various biomolecular systems.

Advancing Molecular Dynamics: Toward Standardization, Integration, and Data Accessibility in Structural Biology

Molecular dynamics (MD) simulations have become a valuable tool in structural biology, offering insights into complex biological systems that are difficult to obtain through experimental techniques alone. The lack of available data sets and structures in most published computational work has limited other researchers' use of these models. In recent years, the emergence of online sharing platforms and MD database initiatives favor the deposition of ensembles and structures to accompany publications, favoring reuse of the data sets. However, the lack of uniform metadata collection, formats, and what data are deposited limits the impact and its use by different communities that are not necessarily experts in MD. This Perspective highlights the need for standardization and better resource sharing for processing and interpreting MD simulation results, akin to efforts in other areas of structural biology. As the field moves forward, we will see an increase in popularity and benefits of MD-based integrative approaches combining experimental data and simulations through probabilistic reasoning, but these too are limited by uniformity in experimental data availability and choices on how the data are modeled that are not trivial to decipher from papers. Other fields have addressed similar challenges comprehensively by establishing task forces with different degrees of success. The large scope and number of communities to represent the breadth of types of MD simulations complicates a parallel approach that would fit all. Thus, each group typically decides what data and which format to upload on servers like Zenodo. Uploading data with FAIR (findable, accessible, interoperable, reusable) principles in mind including optimal metadata collection will make the data more accessible and actionable by the community. Such a wealth of simulation data will foster method development and infrastructure advancements, thus propelling the field forward.

Systematic analysis, aggregation and visualisation of interaction fingerprints for molecular dynamics simulation data

Computational methods such as molecular docking or molecular dynamics (MD) simulations have been developed to simulate and explore the interactions between biomolecules. However, the interactions obtained using these methods are difficult to analyse and evaluate. Interaction fingerprints (IFPs) have been proposed to derive interactions from static 3D coordinates and transform them into 1D bit vectors. More recently, the concept has been applied to derive IFPs from MD simulations, which adds a layer of complexity by adding the temporal motion and dynamics of a system. As a result, many IFPs are obtained from one MD simulation, resulting in a large number of individual IFPs that are difficult to analyse compared to IFPs derived from static 3D structures. Scientific contribution: We introduce a new method to systematically aggregate IFPs derived from MD simulation data. In addition, we propose visualisations to effectively analyse and compare IFPs derived from MD simulation data to account for the temporal evolution of interactions and to compare IFPs across different MD simulations. This has been implemented as a freely available Python library and can therefore be easily adopted by other researchers and to different MD simulation datasets.

All-Atom Biomolecular Simulation in the Exascale Era

Exascale supercomputers have opened the door to dynamic simulations, facilitated by AI/ML techniques, that model biomolecular motions over unprecedented length and time scales. This new capability holds the potential to revolutionize our understanding of fundamental biological processes. Here we report on some of the major advances that were discussed at a recent CECAM workshop in Pisa, Italy, on the topic with a primary focus on atomic-level simulations. First, we highlight examples of current large-scale biomolecular simulations and the future possibilities enabled by crossing the exascale threshold. Next, we discuss challenges to be overcome in optimizing the usage of these powerful resources. Finally, we close by listing several grand challenge problems that could be investigated with this new computer architecture.

The origin and structural evolution of de novo genes in Drosophila

Recent studies reveal that de novo gene origination from previously non-genic sequences is a common mechanism for gene innovation. These young genes provide an opportunity to study the structural and functional origins of proteins. Here, we combine high-quality base-level whole-genome alignments and computational structural modeling to study the origination, evolution, and protein structures of lineage-specific de novo genes. We identify 555 de novo gene candidates in D. melanogaster that originated within the Drosophilinae lineage. Sequence composition, evolutionary rates, and expression patterns indicate possible gradual functional or adaptive shifts with their gene ages. Surprisingly, we find little overall protein structural changes in candidates from the Drosophilinae lineage. We identify several candidates with potentially well-folded protein structures. Ancestral sequence reconstruction analysis reveals that most potentially well-folded candidates are often born well-folded. Single-cell RNA-seq analysis in testis shows that although most de novo gene candidates are enriched in spermatocytes, several young candidates are biased towards the early spermatogenesis stage, indicating potentially important but less emphasized roles of early germline cells in the de novo gene origination in testis. This study provides a systematic overview of the origin, evolution, and protein structural changes of Drosophilinae-specific de novo genes.

Computational approaches to investigate fluoride binding, selectivity and transport across the membrane

The use of molecular dynamics (MD) simulations to study biomolecular systems has proven reliable in elucidating atomic-level details of structure and function. In this chapter, MD simulations were used to uncover new insights into two phylogenetically unrelated bacterial fluoride (F-) exporters: the CLCF F-/H+ antiporter and the Fluc F- channel. The CLCF antiporter, a member of the broader CLC family, has previously revealed unique stoichiometry, anion-coordinating residues, and the absence of an internal glutamate crucial for proton import in the CLCs. Through MD simulations enhanced with umbrella sampling, we provide insights into the energetics and mechanism of the CLCF transport process, including its selectivity for F- over HF. In contrast, the Fluc F- channel presents a novel architecture as a dual topology dimer, featuring two pores for F- export and a central non-transported sodium ion. Using computational electrophysiology, we simulate the electrochemical gradient necessary for F- export in Fluc and reveal details about the coordination and hydration of both F- and the central sodium ion. The procedures described here delineate the specifics of these advanced techniques and can also be adapted to investigate other membrane protein systems.

Probing Mastoparan-like Antimicrobial Peptides Interaction with Model Membrane Through Energy Landscape Analysis

Antimicrobial Peptides (AMPs) have emerged as promising alternatives to conventional antibiotics due to their capacity to disrupt the lipid packing of bacterial cell membranes. This mechanism of action may prevent the development of resistance by bacteria. Understanding their role in lipid packing disruption and their structural properties upon interaction with bacterial membranes is highly desirable. In this study, we employed Molecular Dynamics simulations and the Energy Landscape Visualization Method (ELViM) to characterize and compare the conformational ensembles of mastoparan-like Polybia-MP1 and its analogous H-MP1, in which histidines replace lysine residues. Two situations were analyzed: (i) the peptides in their free state in an aqueous solution containing water and ions and (ii) the peptides spontaneously adsorbing onto an anionic lipid bilayer, used as a bacteria membrane mimetic. ELViM was used to project a single effective conformational phase space for both peptides, providing a comparative analysis. This projection enabled us to map the conformational ensembles of each peptide in an aqueous solution and assess the structural effects of substituting lysines with histidines in H-MP1. Furthermore, a single conformational phase space analysis was employed to describe structural changes during the adsorption process using the same framework. We show that ELViM provides a comprehensive analysis, able to identify discrepancies in the conformational ensembles of these peptides that may affect their affinity to the membrane and adsorption kinetics.

Diffusion Coefficients of Variable-Size Amphiphilic Additives in a Glass-Forming Polyethylene Matrix

Diffusion of additives in polymers is an important issue in the plastics industry since migratory-type molecules are widely used to tune the properties of polymeric composites. Predicting the diffusional behavior of new additives can minimize the need for repetitive experiments. This work presents molecular dynamics simulations at the microsecond time scale and uses the MARTINI force field to estimate self-diffusion coefficients, D, of six monounsaturated amides and their analogs carboxylic acids in polyethylene matrices (PE, MW = 5600 Da). The results are strongly influenced by the glass-forming properties of the PE matrix, which we characterize by three distinct temperatures. The metastability region (T < 325 K), the glass transition temperature (Tg = 256-260 K), and the end of the transition (T ≅ 200 K). Self-diffusion mechanisms are inferred from the results of the dependence of D on the molecular mass of the additive, observing a Rouse-like behavior at high temperatures and deviations from it within the metastability region of the matrix. Interestingly, D values are nonsensitive to the nature of the considered polar head for additives of similar size. The temperature-dependent behavior of D follows, at fixed additive size, a linear Arrhenius pattern at high temperatures and a super Arrhenius trend at lower temperatures, which is well represented with a power law equation as predicted by the Mode Coupling Theory (MCT). We offer a conceptual explanation for the observed super-Arrhenius behavior. This explanation draws on Truhlar and Kohen's interpretation of the available energies at both the initial and the transition states along the diffusion pathway. The matrix's mobility significantly affects solute self-diffusion, yielding equal activation enthalpies for the Arrhenius region or the same power law parameters for the super-Arrhenius regime. Finally, we establish a one-to-one time-equivalence of the self-diffusion processes between CG and all-atom systems for the largest additives and the PE matrix in the high-temperature regime.

Molecular dynamics simulations support a multistep pathway for activation of branched actin filament nucleation by Arp2/3 complex

Actin-related protein 2/3 complex (Arp2/3 complex) catalyzes the nucleation of branched actin filaments that push against membranes in processes like cellular motility and endocytosis. During activation by WASP proteins, the complex must bind WASP and engage the side of a pre-existing (mother) filament before a branched filament is nucleated. Recent high-resolution structures of activated Arp2/3 complex revealed two major sets of activating conformational changes. How these activating conformational changes are triggered by interactions of Arp2/3 complex with actin filaments and WASP remains unclear. Here we use a recent high-resolution structure of Arp2/3 complex at a branch junction to design all-atom molecular dynamics simulations that elucidate the pathway between the active and inactive states. We ran a total of ∼4.6 microseconds of both unbiased and steered all-atom molecular dynamics simulations starting from three different binding states, including Arp2/3 complex within a branch junction, bound only to a mother filament, and alone in solution. These simulations indicate that the contacts with the mother filament are mostly insensitive to the massive rigid body motion that moves Arp2 and Arp3 into a short pitch helical (filament-like) arrangement, suggesting actin filaments alone do not stimulate the short pitch conformational change. In contrast, contacts with the mother filament stabilize subunit flattening in Arp3, an intrasubunit change that converts Arp3 from a conformation that mimics an actin monomer to one that mimics a filamentous actin subunit. Our results support a multistep activation pathway that has important implications for understanding how WASP-mediated activation allows Arp2/3 complex to assemble force-producing actin networks.

Brownian dynamics simulations of mesoscale chromatin fibers

The relationship between chromatin architecture and function defines a central problem in biology and medicine. Many computational chromatin models with atomic, coarse-grained, mesoscale, and polymer resolution have been used to shed light onto the mechanisms that dictate genome folding and regulation of gene expression. The associated simulation techniques range from Monte Carlo to molecular, Brownian, and Langevin dynamics. Here, we present an efficient Compute Unified Device Architecture (CUDA) implementation of Brownian dynamics (BD) to simulate chromatin fibers at the nucleosome resolution with our chromatin mesoscale model. With the CUDA implementation for computer architectures with graphic processing units (GPUs), we significantly accelerate compute-intensive hydrodynamic tensor calculations in the BD simulations by massive parallelization, boosting the performance a hundred-fold compared with central processing unit calculations. We validate our BD simulation approach by reproducing experimental trends on fiber diffusion and structure as a function of salt, linker histone binding, and histone-tail composition, as well as Monte Carlo equilibrium sampling results. Our approach proves to be physically accurate with performance that makes feasible the study of chromatin fibers in the range of kb or hundreds of nucleosomes (small gene). Such simulations are essential to advance the study of biological processes such as gene regulation and aberrant genome-structure related diseases.

Unveiling the interdomain dynamics of type II DNA topoisomerase through all-atom simulations: Implications for understanding its catalytic cycle

Type IIA DNA topoisomerases are complex molecular nanomachines that manage topological states of the DNA molecule in the cell and play a crucial role in cellular processes such as cell division and transcription. They are also established targets of cancer chemotherapy. Starting from the available crystal structure of a fully catalytic topoisomerase IIA homodimer from Saccharomyces cerevisiae, we constructed three states of this molecular motor primarily changing the configurations of the DNA segment bound in the DNA gate and performed μs-long all-atom molecular simulations. A comprehensive analysis revealed a sliding motion within the DNA gate and a teamwork between the N-gate and DNA gate that may be associated with the necessary molecular events that allow passage of the T-segment of DNA. The observed movement of the ATPase dimer relative to the DNA domain was reflected in different interaction patterns between the K-loops of the transducer domain and the B-A-B form of the bound DNA. Based on the obtained results, we mapped simulated configurations to the structures in the proposed catalytic cycle through which type IIA topoisomerases exert their function and discussed the possible transition events. The results extend our understanding of the mechanism of action of type IIA topoisomerases and provide an atomistic interpretation of some of the observed features of these molecular motors.

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.

Differences between Omicron SARS-CoV-2 RBD and other variants in their ability to interact with cell receptors and monoclonal antibodies

SARS-CoV-2 remains a health threat with the continuous emergence of new variants. This work aims to expand the knowledge about the SARS-CoV-2 receptor-binding domain (RBD) interactions with cell receptors and monoclonal antibodies (mAbs). By using constant-pH Monte Carlo simulations, the free energy of interactions between the RBD from different variants and several partners (Angiotensin-Converting Enzyme-2 (ACE2) polymorphisms and various mAbs) were predicted. Computed RBD-ACE2-binding affinities were higher for two ACE2 polymorphisms (rs142984500 and rs4646116) typically found in Europeans which indicates a genetic susceptibility. This is amplified for Omicron (BA.1) and its sublineages BA.2 and BA.3. The antibody landscape was computationally investigated with the largest set of mAbs so far in the literature. From the 32 studied binders, groups of mAbs were identified from weak to strong binding affinities (e.g. S2K146). These mAbs with strong binding capacity and especially their combination are amenable to experimentation and clinical trials because of their high predicted binding affinities and possible neutralization potential for current known virus mutations and a universal coronavirus.Communicated by Ramaswamy H. Sarma.

The origin and structural evolution of de novo genes in Drosophila

Although previously thought to be unlikely, recent studies have shown that de novo gene origination from previously non-genic sequences is a relatively common mechanism for gene innovation in many species and taxa. These young genes provide a unique set of candidates to study the structural and functional origination of proteins. However, our understanding of their protein structures and how these structures originate and evolve are still limited, due to a lack of systematic studies. Here, we combined high-quality base-level whole genome alignments, bioinformatic analysis, and computational structure modeling to study the origination, evolution, and protein structure of lineage-specific de novo genes. We identified 555 de novo gene candidates in D. melanogaster that originated within the Drosophilinae lineage. We found a gradual shift in sequence composition, evolutionary rates, and expression patterns with their gene ages, which indicates possible gradual shifts or adaptations of their functions. Surprisingly, we found little overall protein structural changes for de novo genes in the Drosophilinae lineage. Using Alphafold2, ESMFold, and molecular dynamics, we identified a number of de novo gene candidates with protein products that are potentially well-folded, many of which are more likely to contain transmembrane and signal proteins compared to other annotated protein-coding genes. Using ancestral sequence reconstruction, we found that most potentially well-folded proteins are often born folded. Interestingly, we observed one case where disordered ancestral proteins become ordered within a relatively short evolutionary time. Single-cell RNA-seq analysis in testis showed that although most de novo genes are enriched in spermatocytes, several young de novo genes are biased in the early spermatogenesis stage, indicating potentially important but less emphasized roles of early germline cells in the de novo gene origination in testis. This study provides a systematic overview of the origin, evolution, and structural changes of Drosophilinae-specific de novo genes.

Correlating the differences in the receptor binding domain of SARS-CoV-2 spike variants on their interactions with human ACE2 receptor

Spike glycoprotein of SARS-CoV-2 variants plays a critical role in infection and transmission through its interaction with human angiotensin converting enzyme 2 (hACE2) receptors. Prior findings using molecular docking and biomolecular studies reported varied findings on the difference in the interactions among the spike variants with the hACE2 receptors. Hence, it is a prerequisite to understand these interactions in a more precise manner. To this end, firstly, we performed ELISA with trimeric spike glycoproteins of SARS-CoV-2 variants including Wuhan Hu-1(Wild), Delta, C.1.2 and Omicron. Further, to study the interactions in a more specific manner by mimicking the natural infection, we developed hACE2 receptors expressing HEK-293T cell line, evaluated their binding efficiencies and competitive binding of spike variants with D614G spike pseudotyped virus. In line with the existing findings, we observed that Omicron had higher binding efficiency compared to Delta in both ELISA and Cellular models. Intriguingly, we found that cellular models could differentiate the subtle differences between the closely related C.1.2 and Delta in their binding to hACE2 receptors. Our study using the cellular model provides a precise method to evaluate the binding interactions between spike sub-lineages to hACE2 receptors.

Molecular modelling of the thermophysical properties of fluids: expectations, limitations, gaps and opportunities

This manuscript provides an overview of the current state of the art in terms of the molecular modelling of the thermophysical properties of fluids. It is intended to manage the expectations and serve as guidance to practising physical chemists, chemical physicists and engineers in terms of the scope and accuracy of the more commonly available intermolecular potentials along with the peculiarities of the software and methods employed in molecular simulations while providing insights on the gaps and opportunities available in this field. The discussion is focused around case studies which showcase both the precision and the limitations of frequently used workflows.

Visualizing the Residue Interaction Landscape of Proteins by Temporal Network Embedding

Characterizing the structural dynamics of proteins with heterogeneous conformational landscapes is crucial to understanding complex biomolecular processes. To this end, dimensionality reduction algorithms are used to produce low-dimensional embeddings of the high-dimensional conformational phase space. However, identifying a compact and informative set of input features for the embedding remains an ongoing challenge. Here, we propose to harness the power of Residue Interaction Networks (RINs) and their centrality measures, established tools to provide a graph theoretical view on molecular structure. Specifically, we combine the closeness centrality, which captures global features of the protein conformation at residue-wise resolution, with EncoderMap, a hybrid neural-network autoencoder/multidimensional-scaling like dimensionality reduction algorithm. We find that the resulting low-dimensional embedding is a meaningful visualization of the residue interaction landscape that resolves structural details of the protein behavior while retaining global interpretability. This feature-based graph embedding of temporal protein graphs makes it possible to apply the general descriptive power of RIN formalisms to the analysis of protein simulations of complex processes such as protein folding and multidomain interactions requiring no protein-specific input. We demonstrate this on simulations of the fast folding protein Trp-Cage and the multidomain signaling protein FAT10. Due to its generality and modularity, the presented approach can easily be transferred to other protein systems.

Genome modeling: From chromatin fibers to genes

The intricacies of the 3D hierarchical organization of the genome have been approached by many creative modeling studies. The specific model/simulation technique combination defines and restricts the system and phenomena that can be investigated. We present the latest modeling developments and studies of the genome, involving models ranging from nucleosome systems and small polynucleosome arrays to chromatin fibers in the kb-range, chromosomes, and whole genomes, while emphasizing gene folding from first principles. Clever combinations allow the exploration of many interesting phenomena involved in gene regulation, such as nucleosome structure and dynamics, nucleosome-nucleosome stacking, polynucleosome array folding, protein regulation of chromatin architecture, mechanisms of gene folding, loop formation, compartmentalization, and structural transitions at the chromosome and genome levels. Gene-level modeling with full details on nucleosome positions, epigenetic factors, and protein binding, in particular, can in principle be scaled up to model chromosomes and cells to study fundamental biological regulation.

Data science techniques in biomolecular force field development

Recent advances in data science are impacting the development of classical force fields. Here we review some ideas and techniques from data science that have been used in force field development, including database construction, atom typing, and machine learning potentials. We highlight how new tools such as active learning and automatic differentiation are facilitating the generation of target data and the direct fitting with macroscopic observables. Philosophical changes on how force field models should be built and used are also discussed. It's inspiring that more accurate biomolecular force fields can be developed with the aid of data science techniques.

Regularized by Physics: Graph Neural Network Parametrized Potentials for the Description of Intermolecular Interactions

Simulations of molecular systems using electronic structure methods are still not feasible for many systems of biological importance. As a result, empirical methods such as force fields (FF) have become an established tool for the simulation of large and complex molecular systems. The parametrization of FF is, however, time-consuming and has traditionally been based on experimental data. Recent years have therefore seen increasing efforts to automatize FF parametrization or to replace FF with machine-learning (ML) based potentials. Here, we propose an alternative strategy to parametrize FF, which makes use of ML and gradient-descent based optimization while retaining a functional form founded in physics. Using a predefined functional form is shown to enable interpretability, robustness, and efficient simulations of large systems over long time scales. To demonstrate the strength of the proposed method, a fixed-charge and a polarizable model are trained on ab initio potential-energy surfaces. Given only information about the constituting elements, the molecular topology, and reference potential energies, the models successfully learn to assign atom types and corresponding FF parameters from scratch. The resulting models and parameters are validated on a wide range of experimentally and computationally derived properties of systems including dimers, pure liquids, and molecular crystals.

The dawn of interoperating spatial models in cell biology

Spatial simulations are becoming an increasingly ubiquitous component in the cycle of discovery, experimentation, and communication across the sciences. In cell biology, many researchers share a vision of developing multiscale models that recapitulate observable behaviors spanning from atoms to cells to tissues. For this dream to become a reality, however, simulation technologies must provide a means for integration and interoperability as they advance. Already, the field has developed numerous methods that span scales of length, time, and complexity to create an extensive body of effective simulation approaches, and although these approaches rarely interoperate, they collectively cover a large spectrum of knowledge that future models may handle in a more unified manner. Here, we discuss the importance of making the data, workflows, and outputs of spatial simulations shareable and interoperable; and how democratization could encourage diverse biologists to participate more easily in developing models to advance our understanding of biological systems.

Metamorphic proteins under a computational microscope: Lessons from a fold-switching RfaH protein

Metamorphic proteins constitute unexpected paradigms of the protein folding problem, as their sequences encode two alternative folds, which reversibly interconvert within biologically relevant timescales to trigger different cellular responses. Once considered a rare aberration, metamorphism may be common among proteins that must respond to rapidly changing environments, exemplified by NusG-like proteins, the only transcription factors present in every domain of life. RfaH, a specialized paralog of bacterial NusG, undergoes an all-α to all-β domain switch to activate expression of virulence and conjugation genes in many animal and plant pathogens and is the quintessential example of a metamorphic protein. The dramatic nature of RfaH structural transformation and the richness of its evolutionary history makes for an excellent model for studying how metamorphic proteins switch folds. Here, we summarize the structural and functional evidence that sparked the discovery of RfaH as a metamorphic protein, the experimental and computational approaches that enabled the description of the molecular mechanism and refolding pathways of its structural interconversion, and the ongoing efforts to find signatures and general properties to ultimately describe the protein metamorphome.

Modified Protein-Water Interactions in CHARMM36m for Thermodynamics and Kinetics of Proteins in Dilute and Crowded Solutions

Proper balance between protein-protein and protein-water interactions is vital for atomistic molecular dynamics (MD) simulations of globular proteins as well as intrinsically disordered proteins (IDPs). The overestimation of protein-protein interactions tends to make IDPs more compact than those in experiments. Likewise, multiple proteins in crowded solutions are aggregated with each other too strongly. To optimize the balance, Lennard-Jones (LJ) interactions between protein and water are often increased about 10% (with a scaling parameter, λ = 1.1) from the existing force fields. Here, we explore the optimal scaling parameter of protein-water LJ interactions for CHARMM36m in conjunction with the modified TIP3P water model, by performing enhanced sampling MD simulations of several peptides in dilute solutions and conventional MD simulations of globular proteins in dilute and crowded solutions. In our simulations, 10% increase of protein-water LJ interaction for the CHARMM36m cannot maintain stability of a small helical peptide, (AAQAA)₃ in a dilute solution and only a small modification of protein-water LJ interaction up to the 3% increase (λ = 1.03) is allowed. The modified protein-water interactions are applicable to other peptides and globular proteins in dilute solutions without changing thermodynamic properties from the original CHARMM36m. However, it has a great impact on the diffusive properties of proteins in crowded solutions, avoiding the formation of too sticky protein-protein interactions.

Interrogating the substrate specificity landscape of UvrC reveals novel insights into its non-canonical function

Although it is relatively unexplored, accumulating data highlight the importance of tripartite crosstalk between nucleotide excision repair (NER), DNA replication, and recombination in the maintenance of genome stability; however, elucidating the underlying mechanisms remains challenging. While Escherichia coli uvrA and uvrB can fully complement polAΔ cells in DNA replication, uvrC attenuates this alternative DNA replication pathway, but the exact mechanism by which uvrC suppresses DNA replication is unknown. Furthermore, the identity of bona fide canonical and non-canonical substrates for UvrCs are undefined. Here, we reveal that Mycobacterium tuberculosis UvrC (MtUvrC) strongly binds to, and robustly cleaves, key intermediates of DNA replication/recombination as compared with the model NER substrates. Notably, inactivation of MtUvrC ATPase activity significantly attenuated its endonuclease activity, thus suggesting a causal link between these two functions. We built an in silico model of the interaction of MtUvrC with the Holliday junction (HJ), using a combination of homology modeling, molecular docking, and molecular dynamic simulations. The model predicted residues that were potentially involved in HJ binding. Six of these residues were mutated either singly or in pairs, and the resulting MtUvrC variants were purified and characterized. Among them, residues Glu595 and Arg597 in the helix-hairpin-helix motif were found to be crucial for the interaction between MtUvrC and HJ; consequently, mutations in these residues, or inhibition of ATP hydrolysis, strongly abrogated its DNA-binding and endonuclease activities. Viewed together, these findings expand the substrate specificity landscape of UvrCs and provide crucial mechanistic insights into the interplay between NER and DNA replication/recombination.

Unraveling membrane properties at the organelle-level with LipidDyn

Cellular membranes are formed from different lipids in various amounts and proportions depending on the subcellular localization. The lipid composition of membranes is sensitive to changes in the cellular environment, and its alterations are linked to several diseases. Lipids not only form lipid-lipid interactions but also interact with other biomolecules, including proteins. Molecular dynamics (MD) simulations are a powerful tool to study the properties of cellular membranes and membrane-protein interactions on different timescales and resolutions. Over the last few years, software and hardware for biomolecular simulations have been optimized to routinely run long simulations of large and complex biological systems. On the other hand, high-throughput techniques based on lipidomics provide accurate estimates of the composition of cellular membranes at the level of subcellular compartments. Lipidomic data can be analyzed to design biologically relevant models of membranes for MD simulations. Similar applications easily result in a massive amount of simulation data where the bottleneck becomes the analysis of the data. In this context, we developed LipidDyn, a Python-based pipeline to streamline the analyses of MD simulations of membranes of different compositions. Once the simulations are collected, LipidDyn provides average properties and time series for several membrane properties such as area per lipid, thickness, order parameters, diffusion motions, lipid density, and lipid enrichment/depletion. The calculations exploit parallelization, and the pipeline includes graphical outputs in a publication-ready form. We applied LipidDyn to different case studies to illustrate its potential, including membranes from cellular compartments and transmembrane protein domains. LipidDyn is available free of charge under the GNU General Public License from https://github.com/ELELAB/LipidDyn.

BioExcel Building Blocks Workflows (BioBB-Wfs), an integrated web-based platform for biomolecular simulations

We present BioExcel Building Blocks Workflows, a web-based graphical user interface (GUI) offering access to a collection of transversal pre-configured biomolecular simulation workflows assembled with the BioExcel Building Blocks library. Available workflows include Molecular Dynamics setup, protein-ligand docking, trajectory analyses and small molecule parameterization. Workflows can be launched in the platform or downloaded to be run in the users’ own premises. Remote launching of long executions to user's available High-Performance computers is possible, only requiring configuration of the appropriate access credentials. The web-based graphical user interface offers a high level of interactivity, with integration with the NGL viewer to visualize and check 3D structures, MDsrv to visualize trajectories, and Plotly to explore 2D plots. The server requires no login but is recommended to store the users’ projects and manage sensitive information such as remote credentials. Private projects can be made public and shared with colleagues with a simple URL. The tool will help biomolecular simulation users with the most common and repetitive processes by means of a very intuitive and interactive graphical user interface. The server is accessible at https://mmb.irbbarcelona.org/biobb-wfs.

Finite element modeling of alpha-helices and tropocollagen molecules with reference to the spike of SARS-CoV-2

The newly developed finite element modeling at the atomic scale was used to predict the static and dynamic response of the alpha-helix (AH) and tropocollagen (TC) protein fragments, the main building blocks of the spike of the SARS-CoV-2. The geometry and morphology of the spike's stalk and its connection to the viral envelope were determined from the combination of most recent Molecular Dynamics simulation and images of Cryo-Electron microscope. The stiffness parameters of the covalent bonds in the main chain of the helix were taken from the literature. The AH and TC were modeled using both beam elements (wire model) and shell elements (ribbon model) in finite element analysis to predict their mechanical properties under tension. The asymptotic stiffening features of AH and TC under tensile loading were revealed and compared with a new analytical solution. The mechanical stiffnesses under other loading conditions, including compression, torsion and bending were also predicted numerically and correlated with the results of the existing MD simulations and tests. The mode shapes and natural frequencies of the spike were predicted using the built FE model. The frequencies were shown to be within the safe range of 1-20 MHz routinely used for medical imaging and diagnosis by means of ultrasound. These results provide a solid theoretical basis for using ultrasound to study damaging coronavirus through transient and resonant vibration at large deformations.

Pharmacoproteomics of Brain Barrier Transporters and Substrate Design for the Brain Targeted Drug Delivery

One of the major reasons why central nervous system (CNS)-drug development has been challenging in the past, is the barriers that prevent substances entering from the blood circulation into the brain. These barriers include the blood-brain barrier (BBB), blood-spinal cord barrier (BSCB), blood-cerebrospinal fluid barrier (BCSFB), and blood-arachnoid barrier (BAB), and they differ from each other in their transporter protein expression and function as well as among the species. The quantitative expression profiles of the transporters in the CNS-barriers have been recently revealed, and in this review, it is described how they affect the pharmacokinetics of compounds and how these expression differences can be taken into account in the prediction of brain drug disposition in humans, an approach called pharmacoproteomics. In recent years, also structural biology and computational resources have progressed remarkably, enabling a detailed understanding of the dynamic processes of transporters. Molecular dynamics simulations (MDS) are currently used commonly to reveal the conformational changes of the transporters and to find the interactions between the substrates and the protein during the binding, translocation in the transporter cavity, and release of the substrate on the other side of the membrane. The computational advancements have also aided in the rational design of transporter-utilizing compounds, including prodrugs that can be actively transported without losing potency towards the pharmacological target. In this review, the state-of-art of these approaches will be also discussed to give insights into the transporter-mediated drug delivery to the CNS.

rsRNASP: A residue-separation-based statistical potential for RNA 3D structure evaluation

Knowledge-based statistical potentials have been shown to be rather effective in protein 3-dimensional (3D) structure evaluation and prediction. Recently, several statistical potentials have been developed for RNA 3D structure evaluation, while their performances are either still at a low level for the test datasets from structure prediction models or dependent on the "black-box" process through neural networks. In this work, we have developed an all-atom distance-dependent statistical potential based on residue separation for RNA 3D structure evaluation, namely rsRNASP, which is composed of short- and long-ranged potentials distinguished by residue separation. The extensive examinations against available RNA test datasets show that rsRNASP has apparently higher performance than the existing statistical potentials for the realistic test datasets with large RNAs from structure prediction models, including the newly released RNA-Puzzles dataset, and is comparable to the existing top statistical potentials for the test datasets with small RNAs or near-native decoys. In addition, rsRNASP is superior to RNA3DCNN, a recently developed scoring function through 3D convolutional neural networks. rsRNASP and the relevant databases are available to the public.

Benzophenone-1 induced aberrant proliferation and metastasis of ovarian cancer cells via activated ERα and Wnt/β-catenin signaling pathways

Benzophenone-1 (BP-1) belongs to personal care product-related contaminants of emerging concern and has been recently reported to induce xenoestrogenic effects. However, the underlying mechanisms leading to the activation of target receptors and subsequent various adverse outcomes remain unclear, which is beneficial to safety and health risk assessment of benzophenone-type ultraviolet filters with their widespread occurrence. Herein, we investigated disrupting effects of BP-1 at environmentally relevant concentrations (10^-9-10^-6 M) on estrogen receptor (ER) α-associated signaling pathways. Molecular dynamics simulations together with yeast-based assays revealed the steady binding of BP-1 to ERα ligand binding domain (LBD) and hence the observed agonistic activity. BP-1 triggered interaction between ERα and β-catenin in human SKOV3 ovarian cancer cells and caused translocation of β-catenin from the cytoplasm to the nucleus, leading to aberrant activation of Wnt/β-catenin. BP-1 consequently induced dissemination of SKOV3 via regulating epithelial-mesenchymal transitions (EMT) biomarkers including minimally downregulating ZO-1 gene to 78.0 ± 10.1% and maximally upregulating MMP9 gene to 144.1 ± 29.7% and promoted 1.03-1.83 fold proliferation, migration and invasion of SKOV3. We provide the first evidence that the BP-1 activated ERα triggers crosstalk between ERα and Wnt/β-catenin pathway, leading to the abnormal stimulation and progression of SKOV3 cancer cells.

Molecular Paradigms for Biological Mechanosensing

Many proteins in living cells are subject to mechanical forces, which can be generated internally by molecular machines, or externally, e.g., by pressure gradients. In general, these forces fall in the piconewton range, which is similar in magnitude to forces experienced by a molecule due to thermal fluctuations. While we would naively expect such moderate forces to produce only minimal changes, a wide variety of "mechanosensing" proteins have evolved with functions that are responsive to forces in this regime. The goal of this article is to provide a physical chemistry perspective on protein-based molecular mechanosensing paradigms used in living systems, and how these paradigms can be explored using novel computational methods.

Accelerating COVID-19 Research Using Molecular Dynamics Simulation

The COVID-19 pandemic has emerged as a global medico-socio-economic disaster. Given the lack of effective therapeutics against SARS-CoV-2, scientists are racing to disseminate suggestions for rapidly deployable therapeutic options, including drug repurposing and repositioning strategies. Molecular dynamics (MD) simulations have provided the opportunity to make rational scientific breakthroughs in a time of crisis. Advancements in these technologies in recent years have become an indispensable tool for scientists studying protein structure, function, dynamics, interactions, and drug discovery. Integrating the structural data obtained from high-resolution methods with MD simulations has helped in comprehending the process of infection and pathogenesis, as well as the SARS-CoV-2 maturation in host cells, in a short duration of time. It has also guided us to identify and prioritize drug targets and new chemical entities, and to repurpose drugs. Here, we discuss how MD simulation has been explored by the scientific community to accelerate and guide translational research on SARS-CoV-2 in the past year. We have also considered future research directions for researchers, where MD simulations can help fill the existing gaps in COVID-19 research.

Machine learning-based prediction of survival prognosis in cervical cancer

BACKGROUND

Accurately forecasting the prognosis could improve cervical cancer management, however, the currently used clinical features are difficult to provide enough information. The aim of this study is to improve forecasting capability by developing a miRNAs-based machine learning survival prediction model.

RESULTS

The expression characteristics of miRNAs were chosen as features for model development. The cervical cancer miRNA expression data was obtained from The Cancer Genome Atlas database. Preprocessing, including unquantified data removal, missing value imputation, samples normalization, log transformation, and feature scaling, was performed. In total, 42 survival-related miRNAs were identified by Cox Proportional-Hazards analysis. The patients were optimally clustered into four groups with three different 5-years survival outcome (≥ 90%, ≈ 65%, ≤ 40%) by K-means clustering algorithm base on top 10 survival-related miRNAs. According to the K-means clustering result, a prediction model with high performance was established. The pathways analysis indicated that the miRNAs used play roles involved in the regulation of cancer stem cells.

CONCLUSION

A miRNAs-based machine learning cervical cancer survival prediction model was developed that robustly stratifies cervical cancer patients into high survival rate (5-years survival rate ≥ 90%), moderate survival rate (5-years survival rate ≈ 65%), and low survival rate (5-years survival rate ≤ 40%).

A MULTISCALE VISION-ILLUSTRATIVE APPLICATIONS FROM BIOLOGY TO ENGINEERING

Modeling and simulation have quickly become equivalent pillars of research along with traditional theory and experimentation. The growing realization that most complex phenomena of interest span many orders of spatial and temporal scales has led to an exponential rise in the development and application of multiscale modeling and simulation over the past two decades. In this perspective, the associate editors of the International Journal for Multiscale Computational Engineering and their co-workers illustrate current applications in their respective fields spanning biomolecular structure and dynamics, civil engineering and materials science, computational mechanics, aerospace and mechanical engineering, and more. Such applications are highly tailored, exploit the latest and ever-evolving advances in both computer hardware and software, and contribute significantly to science, technology, and medical challenges in the 21st century.