BitClust: Fast Geometrical Clustering of Long Molecular Dynamics Simulations

Structural diversity in the membrane-bound hIAPP dimer correlated with distinct membrane disruption mechanisms

Amyloid deposits of the human islet amyloid polypeptide (hIAPP) have been identified in 90% of patients with type II diabetes. Cellular membranes accelerate the hIAPP fibrillation, and the integrity of membranes is also disrupted at the same time, leading to the apoptosis of β cells in pancreas. The molecular mechanism of hIAPP-induced membrane disruption, especially during the initial membrane disruption stage, has not been well understood yet. Herein, we carried out extensive all-atom molecular dynamics simulations investigating the hIAPP dimerization process in the anionic POPG membrane, to provide the detailed molecular mechanisms during the initial hIAPP aggregation stage in the membrane environment. Compared to the hIAPP monomer on the membrane, we observed not only an increase of α-helical structures, but also a substantial increase of β-sheet structures upon spontaneous dimerization. Moreover, the random coiled and α-helical dimer structures insert deep into the membrane interior with a few inter-chain contacts at the C-terminal region, while the β-sheet-rich structures reside on the membrane surface accompanied by strong inter-chain hydrophobic interactions. The coexistence of α and β structures constitutes a diverse structural ensemble of the membrane-bound hIAPP dimer. From α-helical to β-sheet structures, the degree of membrane disruption decreases gradually, and thus the membrane damage induced by random coiled and α-helical structures precedes that induced by β-sheet structures. We speculate that insertion of random coiled and α-helical structures contributes to the initial stage of membrane damage, while β-sheet structures on the membrane surface are more involved in the later stage of fibril-induced membrane disruption.

Distant sequence regions of JBP1 contribute to J-DNA binding

Base-J (β-D-glucopyranosyloxymethyluracil) is a modified DNA nucleotide that replaces 1% of thymine in kinetoplastid flagellates. The biosynthesis and maintenance of base-J depends on the base-J-binding protein 1 (JBP1) that has a thymidine hydroxylase domain and a J-DNA-binding domain (JDBD). How the thymidine hydroxylase domain synergizes with the JDBD to hydroxylate thymine in specific genomic sites, maintaining base-J during semi-conservative DNA replication, remains unclear. Here, we present a crystal structure of the JDBD including a previously disordered DNA-contacting loop and use it as starting point for molecular dynamics simulations and computational docking studies to propose recognition models for JDBD binding to J-DNA. These models guided mutagenesis experiments, providing additional data for docking, which reveals a binding mode for JDBD onto J-DNA. This model, together with the crystallographic structure of the TET2 JBP1-homologue in complex with DNA and the AlphaFold model of full-length JBP1, allowed us to hypothesize that the flexible JBP1 N-terminus contributes to DNA-binding, which we confirmed experimentally. Α high-resolution JBP1:J-DNA complex, which must involve conformational changes, would however need to be determined experimentally to further understand this unique underlying molecular mechanism that ensures replication of epigenetic information.

Consequences and opportunities arising due to sparser single-cell RNA-seq datasets

With the number of cells measured in single-cell RNA sequencing (scRNA-seq) datasets increasing exponentially and concurrent increased sparsity due to more zero counts being measured for many genes, we demonstrate here that downstream analyses on binary-based gene expression give similar results as count-based analyses. Moreover, a binary representation scales up to ~ 50-fold more cells that can be analyzed using the same computational resources. We also highlight the possibilities provided by binarized scRNA-seq data. Development of specialized tools for bit-aware implementations of downstream analytical tasks will enable a more fine-grained resolution of biological heterogeneity.

An Expedited Route to Optical and Electronic Properties at Finite Temperature via Unsupervised Learning

Electronic properties and absorption spectra are the grounds to investigate molecular electronic states and their interactions with the environment. Modeling and computations are required for the molecular understanding and design strategies of photo-active materials and sensors. However, the interpretation of such properties demands expensive computations and dealing with the interplay of electronic excited states with the conformational freedom of the chromophores in complex matrices (i.e., solvents, biomolecules, crystals) at finite temperature. Computational protocols combining time dependent density functional theory and ab initio molecular dynamics (MD) have become very powerful in this field, although they require still a large number of computations for a detailed reproduction of electronic properties, such as band shapes. Besides the ongoing research in more traditional computational chemistry fields, data analysis and machine learning methods have been increasingly employed as complementary approaches for efficient data exploration, prediction and model development, starting from the data resulting from MD simulations and electronic structure calculations. In this work, dataset reduction capabilities by unsupervised clustering techniques applied to MD trajectories are proposed and tested for the ab initio modeling of electronic absorption spectra of two challenging case studies: a non-covalent charge-transfer dimer and a ruthenium complex in solution at room temperature. The K-medoids clustering technique is applied and is proven to be able to reduce by ∼100 times the total cost of excited state calculations on an MD sampling with no loss in the accuracy and it also provides an easier understanding of the representative structures (medoids) to be analyzed on the molecular scale.

Computational Approaches to the Rational Design of Tubulin-Targeting Agents

Microtubules are highly dynamic polymers of α,β-tubulin dimers which play an essential role in numerous cellular processes such as cell proliferation and intracellular transport, making them an attractive target for cancer and neurodegeneration research. To date, a large number of known tubulin binders were derived from natural products, while only one was developed by rational structure-based drug design. Several of these tubulin binders show promising in vitro profiles while presenting unacceptable off-target effects when tested in patients. Therefore, there is a continuing demand for the discovery of safer and more efficient tubulin-targeting agents. Since tubulin structural data is readily available, the employment of computer-aided design techniques can be a key element to focus on the relevant chemical space and guide the design process. Due to the high diversity and quantity of structural data available, we compiled here a guide to the accessible tubulin-ligand structures. Furthermore, we review different ligand and structure-based methods recently used for the successful selection and design of new tubulin-targeting agents.

MDSCAN: RMSD-based HDBSCAN clustering of long molecular dynamics

MOTIVATION

The term clustering designates a comprehensive family of unsupervised learning methods allowing to group similar elements into sets called clusters. Geometrical clustering of molecular dynamics (MD) trajectories is a well-established analysis to gain insights into the conformational behavior of simulated systems. However, popular variants collapse when processing relatively long trajectories because of their quadratic memory or time complexity. From the arsenal of clustering algorithms, HDBSCAN stands out as a hierarchical density-based alternative that provides robust differentiation of intimately related elements from noise data. Although a very efficient implementation of this algorithm is available for programming-skilled users (HDBSCAN*), it cannot treat long trajectories under the de facto molecular similarity metric RMSD.

RESULTS

Here, we propose MDSCAN, an HDBSCAN-inspired software specifically conceived for non-programmers users to perform memory-efficient RMSD-based clustering of long MD trajectories. Methodological improvements over the original version include the encoding of trajectories as a particular class of vantage-point tree (decreasing time complexity), and a dual-heap approach to construct a quasi-minimum spanning tree (reducing memory complexity). MDSCAN was able to process a trajectory of 1 million frames using the RMSD metric in about 21 h with <8 GB of RAM, a task that would have taken a similar time but more than 32 TB of RAM with the accelerated HDBSCAN* implementation generally used.

AVAILABILITY AND IMPLEMENTATION

The source code and documentation of MDSCAN are free and publicly available on GitHub (https://github.com/LQCT/MDScan.git) and as a PyPI package (https://pypi.org/project/mdscan/).

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Early Stages of Misfolding of PAP248-286 at two different pH values: An Insight from Molecular Dynamics Simulations

PAP248-286 peptides, which are highly abundant in human semen, aggregate and form amyloid fibrils that enhance HIV infection. Previous experimental studies have shown that the infection-promoting activity of PAP248-286 begins to increase well before amyloid formation takes place and that pH plays a key role in the enhancement of PAP248-286-related infection. Hence, understanding the early stages of misfolding of the PAP2482-86 peptide is crucial. To this end, we have performed 60 independent MD simulations for a total of 24 µs at two different pH values (4.2 and 7.2). Our data shows that early stages of misfolding of the PAP248-286 peptide is a multistage process and that the first step of the process is a transition from an “I-shaped” structure to a “U-shaped” structure. We further observed that the structure of PAP248-286 at the two different pH values shows significantly different features. At pH 4.2, the peptide has less intra-molecular H-bonds and a reduced α-helical content than at pH 7.2. Moreover, differences in intra-peptide residues contacts are also observed at the two pH values. Finally, free energy landscape analysis shows that there are more local minima in the energy surface of the peptide at pH 7.2 than at pH 4.2. Overall, the present study elucidates the early stages of misfolding of the PAP248-286 peptide at the atomic level, thus possibly opening new avenues in structure-based drug discovery against HIV infection.

Identification of Bacterial Membrane Selectivity of Romo1-Derived Antimicrobial Peptide AMPR-22 via Molecular Dynamics

The abuse or misuse of antibiotics has caused the emergence of extensively drug-resistant (XDR) bacteria, rendering most antibiotics ineffective and increasing the mortality rate of patients with bacteremia or sepsis. Antimicrobial peptides (AMPs) are proposed to overcome this problem; however, many AMPs have attenuated antimicrobial activities with hemolytic toxicity in blood. Recently, AMPR-11 and its optimized derivative, AMPR-22, were reported to be potential candidates for the treatment of sepsis with a broad spectrum of antimicrobial activity and low hemolytic toxicity. Here, we performed molecular dynamics (MD) simulations to clarify the mechanism of lower hemolytic toxicity and higher efficacy of AMPR-22 at an atomic level. We found four polar residues in AMPR-11 bound to a model mimicking the bacterial inner/outer membranes preferentially over eukaryotic plasma membrane. AMPR-22 whose polar residues were replaced by lysine showed a 2-fold enhanced binding affinity to the bacterial membrane by interacting with bacterial specific lipids (lipid A or cardiolipin) via hydrogen bonds. The MD simulations were confirmed experimentally in models that partially mimic bacteremia conditions in vitro and ex vivo. The present study demonstrates why AMPR-22 showed low hemolytic toxicity and this approach using an MD simulation would be helpful in the development of AMPs.

Wangzaozin A, a potent novel microtubule stabilizer, targets both the taxane and laulimalide sites on β-tubulin through molecular dynamics simulations

Wangzaozin A, an ent-kaurene diterpenoid isolated from Isodon racemosa (Hemsl) Hara, promotes the polymerization of intracellular microtubules as well as purified tubulin, which is similar to other known microtubule stabilizers. Our pharmacological results showed that wangzaozin A induced G2/M cell cycle arrest and the significant inhibition of cancer cell proliferation. A molecular docking study indicated that wangzaozin A could bind to both the taxane and laulimalide (lau) sites on β-tubulin, which is a novel binding mode that differs from that of known microtubule stabilizers. Furthermore, molecular dynamics simulation and binding free energy calculations demonstrated that wangzaozin A could stably bind to taxane and lau sites simultaneously and form a double-bonded complex. The binding mode of wangzaozin A to the taxane site was more similar to that of epothilone A than paclitaxel. Our results demonstrate that wangzaozin A represents a novel class of microtubule stabilizers, and may serve as a potential microtubule-targeting lead compound for further structural optimization.

RCDPeaks: memory-efficient density peaks clustering of long molecular dynamics

MOTIVATION

Density Peaks is a widely spread clustering algorithm that has been previously applied to Molecular Dynamics (MD) simulations. Its conception of cluster centers as elements displaying both a high density of neighbors and a large distance to other elements of high density, particularly fits the nature of a geometrical converged MD simulation. Despite its theoretical convenience, implementations of Density Peaks carry a quadratic memory complexity that only permits the analysis of relatively short trajectories.

RESULTS

Here, we describe DP+, an exact novel implementation of Density Peaks that drastically reduces the RAM consumption in comparison to the scarcely available alternatives designed for MD. Based on DP+, we developed RCDPeaks, a refined variant of the original Density Peaks algorithm. Through the use of DP+, RCDPeaks was able to cluster a one-million frames trajectory using less than 4.5 GB of RAM, a task that would have taken more than 2 TB and about 3× more time with the fastest and less memory-hunger alternative currently available. Other key features of RCDPeaks include the automatic selection of parameters, the screening of center candidates and the geometrical refining of returned clusters.

AVAILABILITY AND IMPLEMENTATION

The source code and documentation of RCDPeaks are free and publicly available on GitHub (https://github.com/LQCT/RCDPeaks.git).

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Discovery of novel microtubule stabilizers targeting taxane binding site by applying molecular docking, molecular dynamics simulation, and anticancer activity testing

Disruption of the dynamic equilibrium of microtubules can induce cell cycle arrest in G2/M phase and apoptosis. Hence, discovery of novel tubulin polymerization inhibitors is very necessary and an important task in drug research and development for treatment of various tumors. In this investigation, 50 compounds were screened as microtubule stabilizers targeting the taxane site by combination of molecular docking methods. Among these hits, hits 19 and 38 with novel scaffolds exhibited the highest anti-proliferative activity with IC₅₀ ranging from 9.50 to 13.81 μM in four cancer cell lines. The molecular dynamics simulations confirmed that tubulin and two hits could form stable systems. Meanwhile, the mechanism of the interactions between tubulin and two hits at simulated physiological conditions were probed. The in vitro tubulin polymerization assay revealed hits 19 and 38 were able to promote tubulin polymerization in a dose-dependent manner. Further, the immunofluorescence assay suggested that hits 19 and 38 could accelerate microtubule assembly in A549 and HeLa cells. Finally, studies on antitumor activity indicated that hits 19 and 38 induced G2/M phase cell cycle arrest and apoptosis, and inhibited cancer cell motility and migration in A549 and HeLa cells. Importantly, hit38 exhibited better anti-tubulin and anti-cancer activity than hit19 in A549 and HeLa cells. Therefore, these results suggest that hit38 represents a promising microtubule stabilizer for treating cancer and deserves further investigation.

Carbohydrate structure database (CSDB) oligosaccharide conformation tool

Population analysis in terms of glycosidic torsion angles is frequently used to reveal preferred conformers of glycans. However, due to high structural diversity and flexibility of carbohydrates, conformational characterization of complex glycans can be a challenging task. Herein we present a conformation module of oligosaccharide fragments occurring in natural glycan structures developed on the platform of the Carbohydrate Structure Database (CSDB). Currently, this module deposits free energy surface and conformer abundance maps plotted as a function of glycosidic torsions for 194 inter-residue bonds. Data are automatically and continuously derived from explicit-solvent molecular dynamics (MD) simulations. The module was also supplemented with high-temperature MD data of saccharides (2403 maps) provided by GlycoMapsDB (hosted by GLYCOSCIENCES.de project). Conformational data defined by up to four torsional degrees of freedom can be freely explored using a web interface of the module available at http://csdb.glycoscience.ru/database/core/search_conf.html.

BitQT: a graph-based approach to the quality threshold clustering of molecular dynamics

MOTIVATION

Classical Molecular Dynamics (MD) is a standard computational approach to model time-dependent processes at the atomic level. The inherent sparsity of increasingly huge generated trajectories demands clustering algorithms to reduce other post-simulation analysis complexity. The Quality Threshold (QT) variant is an appealing one from the vast number of available clustering methods. It guarantees that all members of a particular cluster will maintain a collective similarity established by a user-defined threshold. Unfortunately, its high computational cost for processing big data limits its application in the molecular simulation field.

RESULTS

In this work, we propose a methodological parallel between QT clustering and another well-known algorithm in the field of Graph Theory, the Maximum Clique Problem. Molecular trajectories are represented as graphs whose nodes designate conformations, while unweighted edges indicate mutual similarity between nodes. The use of a binary-encoded RMSD matrix coupled to the exploitation of bitwise operations to extract clusters significantly contributes to reaching a very affordable algorithm compared to the few implementations of QT for MD available in the literature. Our alternative provides results in good agreement with the exact one while strictly preserving the collective similarity of clusters.

AVAILABILITY AND IMPLEMENTATION

The source code and documentation of BitQT are free and publicly available on GitHub (https://github.com/LQCT/BitQT.git) and ReadTheDocs (https://bitqt.readthedocs.io/en/latest/), respectively.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Discovery of microtubule stabilizers with novel scaffold structures based on virtual screening, biological evaluation, and molecular dynamics simulation

Disrupting the dynamics and structures of microtubules can perturb mitotic spindle formation, cause cell cycle arrest in G2/M phase, and subsequently lead to cellular death via apoptosis. In this investigation, the structure-based virtual screening methods, including molecular docking and rescoring, and similarity analysis of interaction molecular fingerprints, were developed to discover novel tubulin inhibitors from ChemDiv database with 1,601,806 compounds. The screened compounds were further filtered by PAINS, ADME/T, Toxscore, SAscore, and Drug-likeness analysis. Finally, 17 hit compounds were selected, and then submitted to the biologic evaluation. Among these hits, the P2 exhibited the strongest antiproliferative activity against four tumor cells including HeLa, HepG2, MCF-7, and A549. The in vitro tubulin polymerization assay revealed P2 could promote tubulin polymerization in a dose dependent manner. Finally, in order to analyze the interaction modes of complexes, the molecular dynamics simulation was performed to investigate the interactions between P2 and tubulin. The molecular dynamics simulation analysis showed that P2 could stably bind to taxane site, induced H6-H7, B9-B10, and M-loop regions changes. The ΔG_bind energies of tubulin-P2 and tubulin-paclitaxel were -68.25 ± 12.98 and -146.05 ± 16.17 kJ mol^-1, respectively, which were in line with the results of the experimental test. Therefore, P2 has been well characterized as lead compounds for developing new tubulin inhibitors with potential anticancer activity.

quicksom: Self-Organizing Maps on GPUs for clustering of molecular dynamics trajectories

Summary

We implemented the Self-Organizing Maps algorithm running efficiently on GPUs, and also provide several clustering methods of the resulting maps. We provide scripts and a use case to cluster macro-molecular conformations generated by molecular dynamics simulations.

Availability and implementation

The method is available on GitHub and distributed as a pip package.