Development Of Efficient Computational Methods Research Articles

Empowering Graph Neural Networks with Block-Based Dual Adaptive Deep Adjustment for Drug Resistance-Related NcRNA Discovery.

Drug resistance to chemotherapeutic agents remains a formidable challenge in cancer treatment, significantly impacting treatment efficacy. Extensive research has exposed the intimate involvement of noncoding RNAs (ncRNAs) in conferring resistance to cancer drugs. Understanding the intricate associations between ncRNAs and drug resistance is of pivotal importance in advancing clinical interventions and expediting drug development. However, traditional biological experimental methods are hampered by limitations, such as labor intensiveness, time consumption, and constraints in scalability. Addressing these challenges necessitates the development of efficient computational methods for the accurate prediction of potential ncRNA-drug resistance associations (NDRA). However, most existing predictive models primarily focus on known ncRNA-drug resistance associations, often neglecting the critical aspect of similarity information between ncRNAs and drug resistance. This oversight may hinder the accuracy of characterizing these associations. To overcome the limitations of existing computational models, we proposed B-NDRA, a computational framework designed for the discovery of drug resistance-related ncRNA. Initially, we constructed a heterogeneous graph that integrates ncRNA-drug resistance pairs, leveraging both known associations and similarity fusion information between ncRNAs and drug resistance. Subsequently, we employed an attention mechanism to aggregate local features of graph nodes following a dimensionality reduction of node features. Further, a graph neural network (GNN) facilitated the learning of global node embeddings. Notably, the integration of dual adaptive deep adjustment architectures, encompassing intrablock and interblock methodologies, enabled efficient extraction of global features while balancing local and global features. Finally, B-NDRA employed a multilayer perceptron to predict associations between ncRNAs and drug resistance. Through rigorous 5-fold cross-validation, B-NDRA achieved average AUC, AUPR, Accuracy, Precision, Recall, and F1-score values of 92.2%, 91.9%, 84.88%, 86.9%, 82.37%, and 84.44%, respectively. Furthermore, comparative evaluations were conducted on established models, namely, GAEMDA, GRPAMDA, and LRGCPND. The results, obtained through three distinct 5-fold cross-validation strategies, demonstrated a notable performance improvement across almost all metrics for our B-NDRA. Specific case studies targeting Doxorubicin and Imatinib further validated the practicality of our B-NDRA in discovering potential NDRA. These results confirm the potential of our B-NDRA as a valuable tool in advancing cancer research and therapeutic development. The source code and data set of B-NDRA can be found at https://github.com/XuanLi1145/B-NDRA.

Comparing the performance of machine learning methods in predicting soil seed bank persistence

Soil seed banks are pivotal for understanding plant population life histories and elucidating plant-environment relationships. Although experimental exploration of persistent seed banks yields valuable insights, it is often time-consming, labor-intensive, and costly. Consequently, the development of efficient computational methods like machine learning techniques is crucial to augment traditional fieldwork and enhance our capacity for effective seed bank data analysis and interpretation. In this study, we collated a dataset of 398 observations from the TRY database, encompassing measurements such as seed thickness, seed length, seed width, seed dry mass, and seed bank type. Notably, some species' seed bank types within this dataset are identified as having a soil seed bank, without distinguishing between persistent or transient seed banks. Our primary objective was to utilize this data to predict whether a species' soil seed bank is persistent or transient. To this end, we evaluated the performance of four classic machine learning methods—Adaptive Boosting (Adaboost), Artificial Neural Network (ANN), Random Forest (RF), and Support Vector Machine (SVM)—in predicting seed bank types. We gauged the performance of each model using five metrics: Accuracy, AUC (Area Under the Curve), kappa, sensitivity, and specificity. Among the four models examined, the ANN model proved to be the most effective in predicting the persistence of soil seed banks, suggesting its efficacy in predicting seed bank types based on a limited number of independent variables. This study underscores the potential of machine learning methods in soil seed bank research and paves the way for future investigations. It can potentially aid in crafting conservation strategies by enhancing our understanding of plant population dynamics.

QuoteTarget: A sequence-based transformer protein language model to identify potentially druggable protein targets.

The development of efficient computational methods for drug target protein identification can compensate for the high cost of experiments and is therefore of great significance for drug development. However, existing structure-based drug target protein-identification algorithms are limited by the insufficient number of proteins with experimentally resolved structures. Moreover, sequence-based algorithms cannot effectively extract information from protein sequences and thus display insufficient accuracy. Here, we combined the sequence-based self-supervised pretraining protein language model ESM1b with a graph convolutional neural network classifier to develop an improved, sequence-based drug target protein identification method. This complete model, named QuoteTarget, efficiently encodes proteins based on sequence information alone and achieves an accuracy of 95% with the nonredundant drug target and nondrug target datasets constructed for this study. When applied to all proteins from Homo sapiens, QuoteTarget identified 1213 potential undeveloped drug target proteins. We further inferred residue-binding weights from the well-trained network using the gradient-weighted class activation mapping (Grad-Cam) algorithm. Notably, we found that without any binding site information input, significant residues inferred by the model closely match the experimentally confirmed drug molecule-binding sites. Thus, our work provides a highly effective sequence-based identifier for drug target proteins, as well to yield new insights into recognizing drug molecule-binding sites. The entire model is available at https://github.com/Chenjxjx/drug-target-prediction.

PTPAMP: prediction tool for plant-derived antimicrobial peptides.

The emergence of antimicrobial peptides (AMPs) as a potential alternative to conventional antibiotics has led to the development of efficient computational methods for predicting AMPs. Among all organisms, the presence of multiple genes encoding AMPs in plants demands the development of a plant-based prediction tool. To this end, we developed models based on multiple peptide features like amino acid composition, dipeptide composition, and physicochemical attributes for predicting plant-derived AMPs. The selected compositional models are integrated into a web server termed PTPAMP. The designed web server is capable of classifying a query peptide sequence into four functional activities, i.e., antimicrobial (AMP), antibacterial (ABP), antifungal (AFP), and antiviral (AVP). Our models achieved an average area under the curve of 0.95, 0.91, 0.85, and 0.88 for AMP, ABP, AFP, and AVP, respectively, on benchmark datasets, which were ~ 6.75% higher than the state-of-the-art methods. Moreover, our analysis indicates the abundance of cysteine residues in plant-derived AMPs and the distribution of other residues like G, S, K, and R, which differ as per the peptide structural family. Finally, we have developed a user-friendly web server, available at the URL: http://www.nipgr.ac.in/PTPAMP/ . We expect the substantial input of this predictor for high-throughput identification of plant-derived AMPs followed by additional insights into their functions.

An interpretable machine learning method for homo-trimeric protein interface residue-residue interaction prediction

Protein-protein interaction plays an important role in life activities. A more fine-grained analysis, such as residues and atoms level, will better benefit us to understand the mechanism for inter-protein interaction and drug design. The development of efficient computational methods to reduce trials and errors, as well as assisting experimental researchers to determine the complex structure are some of the ongoing studies in the field. The research of trimer protein interface, especially homotrimer, has been rarely studied. In this paper, we proposed an interpretable machine learning method for homo-trimeric protein interface residue pairs prediction. The structure, sequence, and physicochemical information are intergraded as feature input fed to model for training. Graph model is utilized to present spatial information for intra-protein. Matrix factorization captures the different features' interactions. Kernel function is designed to auto-acquire the adjacent information of our target residue pairs. The accuracy rate achieves 54.5% in an independent test set. Sequence and structure alignment exhibit the ability of model self-study. Our model indicates the biological significance between sequence and structure, and could be auxiliary for reducing trials and errors in the fields of protein complex determination and protein-protein docking, etc. SignificanceProtein complex structures are significant for understanding protein function and promising functional protein design. With data increasing, some computational tools have been developed for protein complex residue contact prediction, which is one of the most significant steps for complex structure prediction. But for homo-trimeric protein, the sequence-based deep learning predictors are infeasible for homologous sequences, and the algorithm black box prevents us from understanding of each step operation. In this way, we propose an interpreting machine learning method for homo-trimeric protein interface residue-residue interaction prediction, and the predictor shows a good performance. Our work provides a computational auxiliary way for determining the homo-trimeric proteins interface residue pairs which will be further verified by wet experiments, and and gives a hand for the downstream works, such as protein-protein docking, protein complex structure prediction and drug design.

A review of the scientific contributions by Massimo Salvatores to nuclear reactor physics and engineering

Massimo Salvatores has been an eminent personality of the nuclear science community. He is universally recognized as one of the scientists who helped to maintain research excellence in the nuclear field throughout the past 50 years. His activity covers the theoretical aspects of nuclear reactor physics, the development of efficient computational methods and tools, and experimental analysis. This paper reviews the scientific production of Massimo Salvatores, highlighting what are believed to be his most relevant contributions to nuclear science and engineering.

Baseline Model for Predicting Protein-Ligand Unbinding Kinetics through Machine Learning.

Derivation of structure-kinetics relationships can help rational design and development of new small-molecule drug candidates with desired residence times. Efforts are now being directed toward the development of efficient computational methods. Currently, there is a lack of solid, high-throughput binding kinetics prediction approaches on bigger datasets. We present a prediction method for binding kinetics based on the machine learning analysis of protein-ligand structural features, which can serve as a baseline for more sophisticated methods utilizing molecular dynamics (MD). We showed that the random forest algorithm is capable of learning the protein binding site secondary structure and backbone/side-chain features to predict the binding kinetics of protein-ligand complexes but still with inferior performance to that of MD-based descriptor analysis. MD simulations had been applied to a limited number of targets and a series of ligands in terms of kinetics analysis, and we believe that the developed approach may guide new studies. The method was trained on a newly curated database of 501 protein-ligand unbinding rate constants, which can also be used for testing and training the binding kinetics prediction models.

Chain Model for Carbon Nanotube Bundle under Plane Strain Conditions.

Carbon nanotubes (CNTs) have record high tensile strength and Young’s modulus, which makes them ideal for making super strong yarns, ropes, fillers for composites, solid lubricants, etc. The mechanical properties of CNT bundles have been addressed in a number of experimental and theoretical studies. The development of efficient computational methods for solving this problem is an important step in the design of new CNT-based materials. In the present study, an atomistic chain model is proposed to analyze the mechanical response of CNT bundles under plane strain conditions. The model takes into account the tensile and bending rigidity of the CNT wall, as well as the van der Waals interactions between walls. Due to the discrete character of the model, it is able to describe large curvature of the CNT wall and the fracture of the walls at very high pressures, where both of these problems are difficult to address in frame of continuum mechanics models. As an example, equilibrium structures of CNT crystal under biaxial, strain controlled loading are obtained and their thermal stability is analyzed. The obtained results agree well with previously reported data. In addition, a new equilibrium structure with four SNTs in a translational cell is reported. The model offered here can be applied with great efficiency to the analysis of the mechanical properties of CNT bundles composed of single-walled or multi-walled CNTs under plane strain conditions due to considerable reduction in the number of degrees of freedom.

Transient dynamic analysis of cracked structures with multiple contact pairs using generalized HSNC

The development of efficient computational methods for cracked structures is critical in the fields of civil, mechanical, and aerospace engineering since the influence of cracks on structural dynamics can play an important role in design, prognosis, and health monitoring. The nonlinearity caused by the intermittent contact on the crack surfaces typically excludes the use of fast linear methods, without which the computation of the dynamics of complex cracked structures becomes very challenging. In this paper, an efficient computational scheme for predicting both the transient and steady-state responses of cracked structures with multiple contact pairs is introduced. The new algorithm is referred to as the generalized hybrid symbolic–numeric computational (HSNC) method. The generalized HSNC method extends the original HSNC method, which was recently developed for bilinear systems, to general piecewise-linear nonlinear systems. This work also combines the HSNC with the $$\hbox {X}-\hbox {X}_{r}$$ method, a reduced-order modeling technique for cracked structures, to efficiently predict the dynamics of complex structures with cracks. The generalized HSNC approach is based on the idea that the nonlinear response of a cracked structure with multiple contact pairs can be obtained by combining linear responses of the system in each of its linear states. These linear responses can be symbolically expressed as functions of the initial conditions at starting time points in each time range where the system behaves linearly. The transition time where the system switches from one linear state to another is found using a nonlinear optimization solver with the initial values provided by an incremental search process. The method is able to individually track status of each contact pair; therefore, it can be used to predict the dynamics of the system when the crack surfaces are not completely open or closed. Moreover, both the transient and steady-state responses of complex cracked structures under various forcing conditions can be captured by the new method. The generalized HSNC method provides a flexible computational framework that is several orders of magnitude faster than traditional numerical integration methods. The dynamics of a spring–mass system and cantilever beam models that contain one crack and multiple cracks are investigated using the proposed method.

Probabilistic Performance-Based Optimum Seismic Design of RC Structures Considering Soil–Structure Interaction Effects

In spite of the advances in the development of efficient computational methods, reliability-based design optimization (RBDO) of large-scale structures subjected to seismic loading still requires prohibitive computational effort. Additional difficulties arise for the optimum design of reinforced concrete (RC) structures because they typically feature a wide variety of sections and shapes because of architectural requirements and versatility of construction. In this paper, a modified discrete gravitational search algorithm (MDGSA) is proposed and described, which is paired with a metamodel to minimize the total cost of structures (i.e., the sum of construction and repair costs) under deterministic and probabilistic constraints. MDGSA exploits the crossover and mutation operators used in the differential evolution optimization algorithm to improve the exploitation and exploration of the original gravitational search algorithm (GSA). The competency of the proposed method against the original GSA is demonstrated through performance-based optimum seismic design of an example nine-story RC building subject to both deterministic and probabilistic constraints. Nonlinear soil-structure interaction effects are taken into account in the dynamic finite element analysis of the soil-structure system. The metamodel, which is a weighted least squares support vector machine with the Morlet wavelet kernel function, is used to predict the seismic response of the structure. Annual probabilities of nonperformance are considered as the probabilistic constraints, and the Monte Carlo simulation technique is used to evaluate the seismic reliability of the structure. Results indicate that the proposed MDGSA in conjunction with the metamodel can provide an efficient and effective tool for the RBDO of large-scale structures.

DASPfind: new efficient method to predict drug-target interactions.

BackgroundIdentification of novel drug–target interactions (DTIs) is important for drug discovery. Experimental determination of such DTIs is costly and time consuming, hence it necessitates the development of efficient computational methods for the accurate prediction of potential DTIs. To-date, many computational methods have been proposed for this purpose, but they suffer the drawback of a high rate of false positive predictions.ResultsHere, we developed a novel computational DTI prediction method, DASPfind. DASPfind uses simple paths of particular lengths inferred from a graph that describes DTIs, similarities between drugs, and similarities between the protein targets of drugs. We show that on average, over the four gold standard DTI datasets, DASPfind significantly outperforms other existing methods when the single top-ranked predictions are considered, resulting in 46.17 % of these predictions being correct, and it achieves 49.22 % correct single top ranked predictions when the set of all DTIs for a single drug is tested. Furthermore, we demonstrate that our method is best suited for predicting DTIs in cases of drugs with no known targets or with few known targets. We also show the practical use of DASPfind by generating novel predictions for the Ion Channel dataset and validating them manually.ConclusionsDASPfind is a computational method for finding reliable new interactions between drugs and proteins. We show over six different DTI datasets that DASPfind outperforms other state-of-the-art methods when the single top-ranked predictions are considered, or when a drug with no known targets or with few known targets is considered. We illustrate the usefulness and practicality of DASPfind by predicting novel DTIs for the Ion Channel dataset. The validated predictions suggest that DASPfind can be used as an efficient method to identify correct DTIs, thus reducing the cost of necessary experimental verifications in the process of drug discovery. DASPfind can be accessed online at: http://www.cbrc.kaust.edu.sa/daspfind.Graphical abstractThe conceptual workflow for predicting drug-target interactions using DASPfindElectronic supplementary materialThe online version of this article (doi:10.1186/s13321-016-0128-4) contains supplementary material, which is available to authorized users.

Computation of replicated exercise prices by using positive bases

A procedure is provided for computing the replicated exercise prices of a given portfolio. We highlight a matrix-based framework for analyzing option replication. The new matrix formulation allows the development of efficient computational methods in order to determine the replicated exercise prices of a given portfolio by using the theory of positive bases in vector lattices.

Novel Computational Methods to Design Protein-Protein Interactions

Despite the abundance of structural data, we still cannot accurately predict the structural and energetic changes resulting from mutations at protein interfaces. The inadequacy of current computational approaches to the analysis and design of protein-protein interactions has hampered the development of novel therapeutic and diagnostic agents. In this work, we apply a simple physical model that includes only a minimal set of geometrical constraints, excluded volume, and attractive van der Waals interactions to 1) rank the binding affinity of mutants of tetratricopeptide repeat proteins with their cognate peptides, 2) rank the energetics of binding of small designed proteins to the hydrophobic stem region of the influenza hemagglutinin protein, and 3) predict the stability of T4 lysozyme and staphylococcal nuclease mutants. This work will not only lead to a fundamental understanding of protein-protein interactions, but also to the development of efficient computational methods to rationally design protein interfaces with tunable specificity and affinity, and numerous applications in biomedicine.Funded by: NSF DMR-1006537 and PHY-1019147, the Raymond and Beverly Sack- ler Institute for Biological, Physical and Engineering Sciences, and Howard Hughes Medical Institute (HHMI) International Research Fellowship.

Parallel algorithm for calculating the invariant sets of high-dimensional linear systems under uncertainty

The development of efficient computational methods for synthesizing controls of high-dimensional linear systems is an important problem in theoretical mathematics and its applications. This is especially true for systems with geometrical constraints imposed on the controls and uncertain disturbances. It is well known that the synthesis of target controls under the indicated conditions is based on the construction of weakly invariant sets (reverse reachable sets) generated by the solving equations of the process under study. Methods for constructing such equations and corresponding invariant sets are described, and the computational features for high-dimensional systems are discussed. The approaches proposed are based on the previously developed theory and methods of ellipsoidal approximations of multivalued functions.

Systematic Prediction of cis-Regulatory Elements in the Chlamydomonas reinhardtii Genome Using Comparative Genomics

Chlamydomonas reinhardtii is one of the most important microalgae model organisms and has been widely studied toward the understanding of chloroplast functions and various cellular processes. Further exploitation of C. reinhardtii as a model system to elucidate various molecular mechanisms and pathways requires systematic study of gene regulation. However, there is a general lack of genome-scale gene regulation study, such as global cis-regulatory element (CRE) identification, in C. reinhardtii. Recently, large-scale genomic data in microalgae species have become available, which enable the development of efficient computational methods to systematically identify CREs and characterize their roles in microalgae gene regulation. Here, we performed in silico CRE identification at the whole genome level in C. reinhardtii using a comparative genomics-based method. We predicted a large number of CREs in C. reinhardtii that are consistent with experimentally verified CREs. We also discovered that a large percentage of these CREs form combinations and have the potential to work together for coordinated gene regulation in C. reinhardtii. Multiple lines of evidence from literature, gene transcriptional profiles, and gene annotation resources support our prediction. The predicted CREs will serve, to our knowledge, as the first large-scale collection of CREs in C. reinhardtii to facilitate further experimental study of microalgae gene regulation. The accompanying software tool and the predictions in C. reinhardtii are also made available through a Web-accessible database (http://hulab.ucf.edu/research/projects/Microalgae/sdcre/motifcomb.html).

Bridging the gap between social tagging and semantic annotation: E.D. the Entity Describer

AbstractSemantic annotation enables the development of efficient computational methods for analyzing and interacting with information, thus maximizing its value. With the already substantial and constantly expanding data generation capacity of the life sciences as well as the concomitant increase in the knowledge distributed in scientific articles, new ways to produce semantic annotations of this information are crucial. While automated techniques certainly facilitate the process, manual annotation remains the gold standard in most domains. In this manuscript, we describe a prototype mass-collaborative semantic annotation system that, by distributing the annotation workload across the broad community of biomedical researchers, may help to produce the volume of meaningful annotations needed by modern biomedical science. We present E.D., the Entity Describer, a mashup of the Connotea social tagging system, an index of semantic web-accessible controlled vocabularies, and a new public RDF database for storing social semantic annotations.

Bridging the gap between social tagging and semantic annotation: E.D. the Entity Describer

AbstractSemantic annotation enables the development of efficient computational methods for analyzing and interacting with information, thus maximizing its value. With the already substantial and constantly expanding data generation capacity of the life sciences as well as the concomitant increase in the knowledge distributed in scientific articles, new ways to produce semantic annotations of this information are crucial. While automated techniques certainly facilitate the process, manual annotation remains the gold standard in most domains. In this manuscript, we describe a prototype mass-collaborative semantic annotation system that, by distributing the annotation workload across the broad community of biomedical researchers, may help to produce the volume of meaningful annotations needed by modern biomedical science. We present E.D., the Entity Describer, a mashup of the Connotea social tagging system, an index of semantic web-accessible controlled vocabularies, and a new public RDF database for storing social semantic annotations.

Molecular Lipophilicity in Protein Modeling and Drug Design

Hydrophobic interactions play a key role in the folding and maintenance of the 3-dimensional structure of proteins, as well as in the binding of ligands (e.g. drugs) to protein targets. Therefore, quantitative assessment of spatial hydrophobic (lipophilic) properties of these molecules is indispensable for the development of efficient computational methods in drug design. One possible solution to the problem lies in application of a concept of the 3-dimensional molecular hydrophobicity potential (MHP). The formalism of MHP utilizes a set of atomic physicochemical parameters evaluated from octanol-water partition coefficients (log P) of numerous chemical compounds. It permits detailed assessment of the hydrophobic and/or hydrophilic properties of various parts of molecules and may be useful in analysis of protein-protein and protein-ligand interactions. This review surveys recent applications of MHP-based techniques to a number of biologically relevant tasks. Among them are: (i) Detailed assessment of hydrophobic/hydrophilic organization of proteins; (ii) Application of this data to the modeling of structure, dynamics, and function of globular and membrane proteins, membrane-active peptides, etc. (iii) Employment of the MHP-based criteria in docking simulations for ligands binding to receptors. It is demonstrated that the application of the MHP-based techniques in combination with other molecular modeling tools (e.g. Monte Carlo and molecular dynamics simulations, docking, etc.) permits significant improvement to the standard computational approaches, provides additional important insights into the intimate molecular mechanisms driving protein assembling in water and in biological membranes, and helps in the computer-aided drug discovery process.

Development and implementation of an algorithm for detection of protein complexes in large interaction networks

BackgroundAfter complete sequencing of a number of genomes the focus has now turned to proteomics. Advanced proteomics technologies such as two-hybrid assay, mass spectrometry etc. are producing huge data sets of protein-protein interactions which can be portrayed as networks, and one of the burning issues is to find protein complexes in such networks. The enormous size of protein-protein interaction (PPI) networks warrants development of efficient computational methods for extraction of significant complexes.ResultsThis paper presents an algorithm for detection of protein complexes in large interaction networks. In a PPI network, a node represents a protein and an edge represents an interaction. The input to the algorithm is the associated matrix of an interaction network and the outputs are protein complexes. The complexes are determined by way of finding clusters, i. e. the densely connected regions in the network. We also show and analyze some protein complexes generated by the proposed algorithm from typical PPI networks of Escherichia coli and Saccharomyces cerevisiae. A comparison between a PPI and a random network is also performed in the context of the proposed algorithm.ConclusionThe proposed algorithm makes it possible to detect clusters of proteins in PPI networks which mostly represent molecular biological functional units. Therefore, protein complexes determined solely based on interaction data can help us to predict the functions of proteins, and they are also useful to understand and explain certain biological processes.

Super-paramagnetic clustering of yeast gene expression profiles

High-density DNA arrays, used to monitor gene expression at a genomic scale, have produced vast amounts of information which require the development of efficient computational methods to analyze them. The important first step is to extract the fundamental patterns of gene expression inherent in the data. This paper describes the application of a novel clustering algorithm, super-paramagnetic clustering (SPC) to analysis of gene expression profiles that were generated recently during a study of the yeast cell cycle. SPC was used to organize genes into biologically relevant clusters that are suggestive for their co-regulation. Some of the advantages of SPC are its robustness against noise and initialization, a clear signature of cluster formation and splitting, and an unsupervised self-organized determination of the number of clusters at each resolution. Our analysis revealed interesting correlated behavior of several groups of genes which has not been previously identified.