Language：English   Publishing type：Research paper (scientific journal)   Publisher：American Chemical Society (ACS)  

Some functional proteins undergo conformational changes to expose hidden binding sites when a binding molecule approaches their surface. Such binding sites are called cryptic sites and are important targets for drug discovery. However, it is still difficult to correctly predict cryptic sites. Therefore, we introduce an advanced method, CrypToth, for the precise identification of cryptic sites utilizing the topological data analysis such as persistent homology method. This method integrates topological data analysis and mixed-solvent molecular dynamics (MSMD) simulations. To identify hotspots corresponding to cryptic sites, we conducted MSMD simulations using six probes with different chemical properties: dimethyl ether, benzene, phenol, methyl imidazole, acetonitrile, and ethylene glycol. Subsequently, we applied our topological data analysis method to rank hotspots based on the possibility of harboring cryptic sites. Evaluation of CrypToth using nine target proteins containing well-defined cryptic sites revealed its superior performance compared with recent machine-learning methods. As a result, in seven of nine cases, hotspots associated with cryptic sites were ranked the highest. CrypToth can explore hotspots on the protein surface favorable to ligand binding using MSMD simulations with six different probes and then identify hotspots corresponding to cryptic sites by assessing the protein's conformational variability using the topological data analysis. This synergistic approach facilitates the prediction of cryptic sites with a high accuracy.

DOI： 10.1021/acs.jcim.4c02111

PubMed

researchmap

Language：English   Publishing type：Research paper (scientific journal)   Publisher：MDPI AG  

Cryptic sites, which are transient binding sites that emerge through protein conformational changes upon ligand binding, are valuable targets for drug discovery, particularly for allosteric modulators. However, identifying these sites remains challenging because they are often discovered serendipitously when both ligand-binding (holo) and ligand-free (apo) states are experimentally determined. Here, we introduce CrypTothML, a novel framework that integrates mixed-solvent molecular dynamics (MSMD) simulations and machine learning to predict cryptic sites accurately. CrypTothML first identifies hotspots through MSMD simulations using six chemically diverse probes (benzene, dimethyl-ether, phenol, methyl-imidazole, acetonitrile, and ethylene glycol). A machine learning model then ranks these hotspots based on their likelihood of being cryptic sites, incorporating both hotspot-derived and protein-specific features. Evaluation on a curated dataset demonstrated that CrypTothML outperforms recent machine learning-based methods, achieving an AUC-ROC of 0.88 and successfully identifying cryptic sites missed by other methods. Additionally, CrypTothML ranked cryptic sites as the top prediction more frequently than existing methods. This approach provides a powerful strategy for accelerating drug discovery and designing allosteric drugs.

DOI： 10.3390/ijms26104710

PubMed

researchmap

Language：English   Publishing type：Research paper (scientific journal)   Publisher：Oxford University Press (OUP)  

Abstract

Cyclic peptides are versatile therapeutic agents that boast high binding affinity, minimal toxicity, and the potential to engage challenging protein targets. However, the pharmaceutical utility of cyclic peptides is limited by their low membrane permeability—an essential indicator of oral bioavailability and intracellular targeting. Current machine learning-based models of cyclic peptide permeability show variable performance owing to the limitations of experimental data. Furthermore, these methods use features derived from the whole molecule that have traditionally been used to predict small molecules and ignore the unique structural properties of cyclic peptides. This study presents CycPeptMP: an accurate and efficient method to predict cyclic peptide membrane permeability. We designed features for cyclic peptides at the atom-, monomer-, and peptide-levels and seamlessly integrated these into a fusion model using deep learning technology. Additionally, we applied various data augmentation techniques to enhance model training efficiency using the latest data. The fusion model exhibited excellent prediction performance for the logarithm of permeability, with a mean absolute error of $0.355$ and correlation coefficient of $0.883$. Ablation studies demonstrated that all feature levels contributed and were relatively essential to predicting membrane permeability, confirming the effectiveness of augmentation to improve prediction accuracy. A comparison with a molecular dynamics-based method showed that CycPeptMP accurately predicted peptide permeability, which is otherwise difficult to predict using simulations.

DOI： 10.1093/bib/bbae417

PubMed

researchmap

Language：English   Publisher：American Chemical Society (ACS)  

DOI： 10.1021/acs.jcim.2c01573

PubMed

J-GLOBAL

researchmap

Language：English   Publishing type：Research paper (scientific journal)   Publisher：American Chemical Society (ACS)  

Cyclic peptides have attracted attention as a promising pharmaceutical modality due to their potential to selectively inhibit previously undruggable targets, such as intracellular protein-protein interactions. Poor membrane permeability is the biggest bottleneck hindering successful drug discovery based on cyclic peptides. Therefore, the development of computational methods that can predict membrane permeability and support elucidation of the membrane permeation mechanism of drug candidate peptides is much sought after. In this study, we developed a protocol to simulate the behavior in membrane permeation steps and estimate the membrane permeability of large cyclic peptides with more than or equal to 10 residues. This protocol requires the use of a more realistic membrane model than a single-lipid phospholipid bilayer. To select a membrane model, we first analyzed the effect of cholesterol concentration in the model membrane on the potential of mean force and hydrogen bonding networks along the direction perpendicular to the membrane surface as predicted by molecular dynamics simulations using cyclosporine A. These results suggest that a membrane model with 40 or 50 mol % cholesterol was suitable for predicting the permeation process. Subsequently, two types of membrane models containing 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine and 40 and 50 mol % cholesterol were used. To validate the efficiency of our protocol, the membrane permeability of 18 ten-residue peptides was predicted. Correlation coefficients of R > 0.8 between the experimental and calculated permeability values were obtained with both model membranes. The results of this study demonstrate that the lipid membrane is not just a medium but also among the main factors determining the membrane permeability of molecules. The computational protocol proposed in this study and the findings obtained on the effect of membrane model composition will contribute to building a schematic view of the membrane permeation process. Furthermore, the results of this study will eventually aid the elucidation of design rules for peptide drugs with high membrane permeability.

DOI： 10.1021/acs.jcim.2c00931

PubMed

researchmap

Authorship：Lead author   Language：English   Publishing type：Research paper (scientific journal)   Publisher：MDPI AG  

To ensure efficiency in discovery and development, the application of computational technology is essential. Although virtual screening techniques are widely applied in the early stages of drug discovery research, the computational methods used in lead optimization to improve activity and reduce the toxicity of compounds are still evolving. In this study, we propose a method to construct the residue interaction profile of the chemical structure used in the lead optimization by performing “inverse” mixed-solvent molecular dynamics (MSMD) simulation. Contrary to constructing a protein-based, atom interaction profile, we constructed a probe-based, protein residue interaction profile using MSMD trajectories. It provides us the profile of the preferred protein environments of probes without co-crystallized structures. We assessed the method using three probes: benzamidine, catechol, and benzene. As a result, the residue interaction profile of each probe obtained by MSMD was a reasonable physicochemical description of the general non-covalent interaction. Moreover, comparison with the X-ray structure containing each probe as a ligand shows that the map of the interaction profile matches the arrangement of amino acid residues in the X-ray structure.

DOI： 10.3390/ijms23094749

PubMed

researchmap

Language：English   Publishing type：Research paper (scientific journal)   Publisher：MDPI AG  

In the polyomino puzzle, the aim is to fill a finite space using several polyomino pieces with no overlaps or blanks. Because it is an NP-complete combinatorial optimization problem, various probabilistic and approximated approaches have been applied to find solutions. Several previous studies embedded the polyomino puzzle in a QUBO problem, where the original objective function and constraints are transformed into the Hamiltonian function of the simulated Ising model. A solution to the puzzle is obtained by searching for a ground state of Hamiltonian by simulating the dynamics of the multiple-spin system. However, previous methods could solve only tiny polyomino puzzles considering a few combinations because their Hamiltonian designs were not efficient. We propose an improved Hamiltonian design that introduces new constraints and guiding terms to weakly encourage favorable spins and pairs in the early stages of computation. The proposed model solves the pentomino puzzle represented by approximately 2000 spins with >90% probability. Additionally, we extended the method to a generalized problem where each polyomino piece could be used zero or more times and solved it with approximately 100% probability. The proposed method also appeared to be effective for the 3D polycube puzzle, which is similar to applications in fragment-based drug discovery.

DOI： 10.3390/e24030354

PubMed

researchmap

Language：English   Publishing type：Research paper (scientific journal)   Publisher：Oxford University Press (OUP)  

Abstract

Motivation

In recent years, cyclic peptide drugs have been receiving increasing attention because they can target proteins that are difficult to be tackled by conventional small-molecule drugs or antibody drugs. Plasma protein binding rate (%PPB) is a significant pharmacokinetic property of a compound in drug discovery and design. However, due to structural differences, previous computational prediction methods developed for small-molecule compounds cannot be successfully applied to cyclic peptides, and methods for predicting the PPB rate of cyclic peptides with high accuracy are not yet available.

Results

Cyclic peptides are larger than small molecules, and their local structures have a considerable impact on PPB; thus, molecular descriptors expressing residue-level local features of cyclic peptides, instead of those expressing the entire molecule, as well as the circularity of the cyclic peptides should be considered. Therefore, we developed a prediction method named CycPeptPPB using deep learning that considers both factors. First, the macrocycle ring of cyclic peptides was decomposed residue by residue. The residue-based descriptors were arranged according to the sequence information of the cyclic peptide. Furthermore, the circular data augmentation method was used, and the circular convolution method CyclicConv was devised to express the cyclic structure. CycPeptPPB exhibited excellent performance, with mean absolute error (MAE) of 4.79% and correlation coefficient (R) of 0.92 for the public drug dataset, compared to the prediction performance of the existing PPB rate prediction software (MAE=15.08%, R=0.63).

Availability and implementation

The data underlying this article are available in the online supplementary material. The source code of CycPeptPPB is available at https://github.com/akiyamalab/cycpeptppb.

Supplementary information

Supplementary data are available at Bioinformatics online.

DOI： 10.1093/bioinformatics/btab726

PubMed

researchmap

Other Link： https://academic.oup.com/bioinformatics/article-pdf/38/4/1110/49008780/btab726.pdf

Language：Japanese   Publishing type：Research paper (scientific journal)  

DOI： 10.11234/jsbibr.2021.9

J-GLOBAL

researchmap

Language：English   Publishing type：Research paper (scientific journal)  

DOI： 10.3390/genes12091455

Web of Science

PubMed

researchmap

Language：English   Publisher：American Chemical Society (ACS)  

Membrane permeability is a significant obstacle facing the development of cyclic peptide drugs. However, membrane permeation mechanisms are poorly understood. To investigate common features of permeable (and nonpermeable) designs, it is necessary to reproduce the membrane permeation process of cyclic peptides through the lipid bilayer. We simulated the membrane permeation process of 100 six-residue cyclic peptides across the lipid bilayer based on steered molecular dynamics (MD) and replica-exchange umbrella sampling simulations and predicted membrane permeability using the inhomogeneous solubility-diffusion model and a modified version of it. Furthermore, we confirmed the effectiveness of this protocol by predicting the membrane permeability of 56 eight-residue cyclic peptides with diverse chemical structures, including some confidential designs from a pharmaceutical company. As a result, a reasonable correlation between experimentally assessed and calculated membrane permeability of cyclic peptides was observed for the peptide libraries, except for strongly hydrophobic peptides. Our analysis of the MD trajectory demonstrated that most peptides were stabilized in the boundary region between bulk water and membrane and that for most peptides, the process of crossing the center of the membrane is the main obstacle to membrane permeation. The height of this barrier is well correlated with the electrostatic interaction between the peptide and the surrounding media. The structural and energetic features of the representative peptide at each vertical position within the membrane were also analyzed, revealing that peptides permeate the membrane by changing their orientation and conformation according to the surrounding environment.

DOI： 10.1021/acs.jcim.1c00380

PubMed

J-GLOBAL

researchmap

Language：English  

J-GLOBAL

researchmap

Authorship：Lead author,　Corresponding author   Language：English   Publishing type：Research paper (scientific journal)   Publisher：American Chemical Society (ACS)  

Cosolvent molecular dynamics (CMD) simulations involve an MD simulation of a protein in the presence of explicit water molecules mixed with cosolvent molecules to perform hotspot detection, binding site identification, and binding energy estimation, while other existing methods (e.g., MixMD, SILCS, and MDmix) utilize small molecules that represent functional groups of compounds. However, the cosolvent selections employed in these methods differ and there are only a few cosolvents that are commonly used in these methods. In this study, we proposed a systematic method for constructing a set of cosolvents for drug discovery, termed the EXtended PRObes set construction by REpresentative Retrieval (EXPRORER). First, we extracted typical substructures from FDA-approved drugs, generated 138 cosolvent structures, and for each cosolvent molecule, we conducted CMD simulations to generate a spatial probability distribution map of cosolvent atoms (PMAP). Analyses of PMAP similarity revealed that a cosolvent pair with a PMAP similarity greater than 0.70-0.75 shared similar structural features. We present a method for the construction of a cosolvent subset that satisfies a similarity threshold for all cosolvents, and we tested the constructed sets for four proteins. To our knowledge, this is the first study to include a systematic proposal for cosolvent set construction, and thus, the EXPRORER cosolvents will provide deeper insights into ligand binding sites of various proteins.

DOI： 10.1021/acs.jcim.1c00134

PubMed

researchmap

Language：English  

DOI： 10.48550/arXiv.1907.01103

J-GLOBAL

researchmap

Other Link： https://dblp.uni-trier.de/db/journals/corr/corr1907.html#abs-1907-01103

Language：English   Publishing type：Research paper (scientific journal)   Publisher：Springer Science and Business Media LLC  

BACKGROUND: Cyclic peptide-based drug discovery is attracting increasing interest owing to its potential to avoid target protein depletion. In drug discovery, it is important to maintain the biostability of a drug within the proper range. Plasma protein binding (PPB) is the most important index of biostability, and developing a computational method to predict PPB of drug candidate compounds contributes to the acceleration of drug discovery research. PPB prediction of small molecule drug compounds using machine learning has been conducted thus far; however, no study has investigated cyclic peptides because experimental information of cyclic peptides is scarce. RESULTS: First, we adopted sparse modeling and small molecule information to construct a PPB prediction model for cyclic peptides. As cyclic peptide data are limited, applying multidimensional nonlinear models involves concerns regarding overfitting. However, models constructed by sparse modeling can avoid overfitting, offering high generalization performance and interpretability. More than 1000 PPB data of small molecules are available, and we used them to construct a prediction models with two enumeration methods: enumerating lasso solutions (ELS) and forward beam search (FBS). The accuracies of the prediction models constructed by ELS and FBS were equal to or better than those of conventional non-linear models (MAE = 0.167-0.174) on cross-validation of a small molecule compound dataset. Moreover, we showed that the prediction accuracies for cyclic peptides were close to those for small molecule compounds (MAE = 0.194-0.288). Such high accuracy could not be obtained by a simple method of learning from cyclic peptide data directly by lasso regression (MAE = 0.286-0.671) or ridge regression (MAE = 0.244-0.354). CONCLUSION: In this study, we proposed a machine learning techniques that uses low-dimensional sparse modeling to predict the PPB value of cyclic peptides computationally. The low-dimensional sparse model not only exhibits excellent generalization performance but also improves interpretation of the prediction model. This can provide common an noteworthy knowledge for future cyclic peptide drug discovery studies.

DOI： 10.1186/s12859-018-2529-z

PubMed

researchmap

Other Link： http://link.springer.com/article/10.1186/s12859-018-2529-z/fulltext.html

Language：English   Publishing type：Research paper (scientific journal)   Publisher：Springer Science and Business Media LLC  

Druglikeness is a useful concept for screening drug candidate compounds. We developed QEX, which is a new druglikeness index specific to individual targets. QEX is an improvement of the quantitative estimate of druglikeness (QED) method, which is a popular quantitative evaluation method of druglikeness proposed by Bickerton et al. QEX models the physicochemical properties of compounds that act on each target protein based on the concept of QED modeling physicochemical properties from information on US Food and Drug Administration-approved drugs. The result of the evaluation of PubChem assay data revealed that QEX showed better performance than the original QED did (the area under the curve value of the receiver operating characteristic curve improved by 0.069-0.236). We also present the c-Src inhibitor filtering results of the QEX constructed using Src family kinase inhibitors as a case study. QEX distinguished the inhibitors and non-inhibitors better than QED did. QEX works efficiently even when datasets of inactive compounds are unavailable. If both active and inactive compounds are present, QEX can be used as an initial filter to enhance the screening ability of conventional ligand-based virtual screenings.

DOI： 10.1007/s11030-018-9842-3

PubMed

researchmap

Other Link： http://link.springer.com/content/pdf/10.1007/s11030-018-9842-3.pdf

Language：English   Publishing type：Research paper (scientific journal)   Publisher：Elsevier BV  

The need to accelerate large-scale protein-ligand docking in virtual screening against a huge compound database led researchers to propose a strategy that entails memorizing the evaluation result of the partial structure of a compound and reusing it to evaluate other compounds. However, the previous method required frequent disk accesses, resulting in insufficient acceleration. Thus, more efficient memory usage can be expected to lead to further acceleration, and optimal memory usage could be achieved by solving the minimum cost flow problem. In this research, we propose a fast algorithm for the minimum cost flow problem utilizing the characteristics of the graph generated for this problem as constraints. The proposed algorithm, which optimized memory usage, was approximately seven times faster compared to existing minimum cost flow algorithms.

DOI： 10.1016/j.compbiolchem.2018.03.013

PubMed

researchmap

Other Link： https://dblp.uni-trier.de/db/journals/candc/candc74.html#YanagisawaKKOA18

Language：English   Publishing type：Research paper (scientific journal)   Publisher：Springer Science and Business Media LLC  

DOI： 10.1186/s12859-018-2073-x

Scopus

PubMed

researchmap

Language：English   Publishing type：Research paper (scientific journal)   Publisher：Springer Science and Business Media LLC  

Abstract

We propose a new iterative screening contest method to identify target protein inhibitors. After conducting a compound screening contest in 2014, we report results acquired from a contest held in 2015 in this study. Our aims were to identify target enzyme inhibitors and to benchmark a variety of computer-aided drug discovery methods under identical experimental conditions. In both contests, we employed the tyrosine-protein kinase Yes as an example target protein. Participating groups virtually screened possible inhibitors from a library containing 2.4 million compounds. Compounds were ranked based on functional scores obtained using their respective methods, and the top 181 compounds from each group were selected. Our results from the 2015 contest show an improved hit rate when compared to results from the 2014 contest. In addition, we have successfully identified a statistically-warranted method for identifying target inhibitors. Quantitative analysis of the most successful method gave additional insights into important characteristics of the method used.

DOI： 10.1038/s41598-017-10275-4

PubMed

researchmap

Authorship：Lead author   Language：English   Publishing type：Research paper (scientific journal)   Publisher：Oxford University Press (OUP)  

Abstract

Motivation

Recently, the number of available protein tertiary structures and compounds has increased. However, structure-based virtual screening is computationally expensive owing to docking simulations. Thus, methods that filter out obviously unnecessary compounds prior to computationally expensive docking simulations have been proposed. However, the calculation speed of these methods is not fast enough to evaluate ≥ 10 million compounds.

Results

In this article, we propose a novel, docking-based pre-screening protocol named Spresso (Speedy PRE-Screening method with Segmented cOmpounds). Partial structures (fragments) are common among many compounds; therefore, the number of fragment variations needed for evaluation is smaller than that of compounds. Our method increases calculation speeds by ∼200-fold compared to conventional methods.

Availability and Implementation

Spresso is written in C ++ and Python, and is available as an open-source code (http://www.bi.cs.titech.ac.jp/spresso/) under the GPLv3 license.

Supplementary information

Supplementary data are available at Bioinformatics online.

DOI： 10.1093/bioinformatics/btx178

PubMed

researchmap

Language：English   Publishing type：Research paper (scientific journal)   Publisher：Springer Science and Business Media LLC  

Abstract

A search of broader range of chemical space is important for drug discovery. Different methods of computer-aided drug discovery (CADD) are known to propose compounds in different chemical spaces as hit molecules for the same target protein. This study aimed at using multiple CADD methods through open innovation to achieve a level of hit molecule diversity that is not achievable with any particular single method. We held a compound proposal contest, in which multiple research groups participated and predicted inhibitors of tyrosine-protein kinase Yes. This showed whether collective knowledge based on individual approaches helped to obtain hit compounds from a broad range of chemical space and whether the contest-based approach was effective.

DOI： 10.1038/srep17209

PubMed

researchmap

Authorship：Lead author   Language：English   Publishing type：Research paper (scientific journal)   Publisher：Information Processing Society of Japan  

DOI： 10.2197/ipsjtbio.8.21

J-GLOBAL

researchmap

J-GLOBAL

researchmap

J-GLOBAL

researchmap

Authorship：Lead author,　Last author,　Corresponding author   Language：Japanese   Publishing type：Meeting report   Publisher：情報処理学会  

DOI： 10.20729/0002003560

CiNii Books

CiNii Research

researchmap

Language：Japanese   Publishing type：Research paper, summary (national, other academic conference)  

J-GLOBAL

researchmap

Language：Japanese   Publishing type：Research paper, summary (national, other academic conference)  

J-GLOBAL

researchmap

Language：Japanese   Publishing type：Research paper, summary (national, other academic conference)  

J-GLOBAL

researchmap

Language：Japanese   Publishing type：Research paper, summary (national, other academic conference)  

J-GLOBAL

researchmap

Language：Japanese   Publishing type：Research paper, summary (national, other academic conference)   Publisher：情報処理学会  

J-GLOBAL

researchmap

J-GLOBAL

researchmap

J-GLOBAL

researchmap

J-GLOBAL

researchmap

J-GLOBAL

researchmap

J-GLOBAL

researchmap

Language：Japanese   Publishing type：Research paper, summary (national, other academic conference)  

J-GLOBAL

researchmap

Language：Japanese   Publishing type：Research paper, summary (national, other academic conference)  

J-GLOBAL

researchmap

Language：Japanese   Publishing type：Research paper, summary (national, other academic conference)  

J-GLOBAL

researchmap

Language：Japanese   Publishing type：Research paper, summary (national, other academic conference)   Publisher：情報処理学会  

J-GLOBAL

researchmap

Language：Japanese   Publishing type：Research paper, summary (national, other academic conference)   Publisher：情報処理学会  

J-GLOBAL

researchmap

J-GLOBAL

researchmap

J-GLOBAL

researchmap

J-GLOBAL

researchmap

Language：Japanese   Publishing type：Research paper, summary (national, other academic conference)   Publisher：情報処理学会  

J-GLOBAL

researchmap

Language：Japanese   Publishing type：Research paper, summary (national, other academic conference)   Publisher：情報処理学会  

J-GLOBAL

researchmap

J-GLOBAL

researchmap

Language：Japanese   Publishing type：Research paper, summary (national, other academic conference)   Publisher：情報処理学会  

近年注目されている中分子医薬品のうちの 1 つである環状ペプチド医薬品の開発によって，従来の医薬品では狙うことが難しかった細胞内のタンパク質間相互作用（PPI）の阻害などが可能になると期待されている．しかし，一般的に環状ペプチドは細胞膜透過性が低く，経口投与や細胞内標的の阻害が可能な膜透過性の持つ環状ペプチドを選別する必要がある．そこで，充分な膜透過性を持つ環状ペプチドを選別するために，環状ペプチドの膜透過性の予測法の確立が必要とされている．本研究では，環状ペプチドの脂質二重膜の透過過程を分子動力学シミュレーションにて再現し，その軌跡データから膜透過性の高さと相関のある特徴量の抽出を試みた．サンプリングには REST/REUS 法を用いて，広範なコンフォメーションを探索した．また，計算した特徴量を用いて機械学習による膜透過性予測も行った．結果として，10 残基環状ペプチドを用いた単回帰分析では，膜の界面付近での，ペプチドとその周りの分子との静電相互作用が最も膜透過性の実験値と相関があった（r=0.89）．また，6 残基環状ペプチドを用いた機械学習による膜透過性予測においても，膜の界面付近の特徴量が重要であることが示唆された．

J-GLOBAL

researchmap

Language：Japanese   Publishing type：Research paper, summary (national, other academic conference)  

In the development of gapmer ASOs, there is a problem that not only the on-target gene but also off-target genes are down-regulated. Evaluating the off-target effects by experimental screening tests is time-consuming and economically burdensome. Therefore, computational risk assessment is needed. In this study, we proposed a risk score in terms of cleavage by RNase H1, binding energy between an ASO and an mRNA, and the secondary structure of mRNAs. For the evaluation, we used the 13-mer gapmer targeting human APOB (gap-A13) and the 13-mer gapmer targeting human PCSK9 (gap-P13) used in previous study. For each gapmer ASO, Pearson correlation coefficients between the experimental values of log fold-change (logFC) and the assessed risk scores were calculated for a large number of genes. The results of the calculations were r = -0.091 for gap-A13 and r = -0.124 for gap-P13. This result indicates that the risk score is not practical as a valid risk measure. However, there was a statistically significant difference in gene expression intensity before the introduction of ASO between the gene groups with overestimated risk scores and the others. This suggests that the gene expression intensity before ASO introduction is important for evaluating off-target effects.

J-GLOBAL

researchmap

Language：Japanese   Publishing type：Research paper, summary (national, other academic conference)  

The binding energy between nucleic acid sequences is an important factor in considering the binding stability of nucleic acids. Although various methods exist, no method is dedicated for short sequences. The binding energy estimation of short sequences is needed to develop antisense oligonucleotide (ASO) drugs, which has a gap region of approximately 10 bases. In this study, we propose a method for short sequences that enumerate sequences and search through target sequences. We compared to RIsearch, a well-known tool, and the proposed method is 60-90 times faster than the previous tool when the query sequence length is 10-14 nucleotides. In addition, the amount of memory usage was 1.4-116.6MB, indicating that the proposed method can be implemented with a personal computer.

J-GLOBAL

researchmap

Language：Japanese   Publishing type：Research paper, summary (national, other academic conference)   Publisher：情報処理学会  

J-GLOBAL

researchmap

Language：Japanese   Publishing type：Research paper, summary (national, other academic conference)  

J-GLOBAL

researchmap

Language：Japanese   Publishing type：Research paper, summary (national, other academic conference)   Publisher：情報処理学会  

近年，フラグメントと呼ばれる分子量が小さい分子をもとにリード化合物を設計する Fragment-Based Drug Design (FBDD) の方法論が注目されている．なかでも，離れたポケットを占めるフラグメント同士を結合させる Fragment Linking は，複数の結合部位を有効に活用できる利点があるものの，他の FBDD 手法と比べて難しいことが指摘されている．また，距離のあるリンキングへ対応できる汎用的な手法は知られていない．本研究では，長距離のリンキングを可能にするため，リンカーの部品となるリンカー要素を集めたライブラリを予め作成し，その中から複数のリンカー要素を逐次接続し，立体的な配置を探索するフラグメントリンキング手法を提案する．また，リンカー選択に用いる評価関数として，標的とのドッキングスコアを組み込んだ複数の関数を比較検討を行った．その結果，例題として用いたタンパク質チロシンホスファターゼ 1B を標的とした阻害剤設計実験では最大 10.4 Å 離れたフラグメント間のリンキングに成功した．また，リンカー要素の選択に用いる評価関数を 6 種類設計し，それらの効果を検証した．

J-GLOBAL

researchmap

Language：Japanese   Publishing type：Research paper, summary (national, other academic conference)  

In the development of antisense oligonucleotide (ASO) drugs, a large number of ASOs that are reverse complementary to the target transcripts are designed with 10-20 bases. ASOs with high affinity are then experimentally determined. Since wet experiment is time-consuming and costly, computational prediction of desired ASOs is highly demanded. In this study, we analyzed the relationshps between the affinity of ASO, and inter- and intra-hybridization energies of ASO and/or mRNA. As a result, the binding energy between ASO and mRNA was the most correlated with the inhibition rate of gene expression. In addition, the inhibition rate tends to be low, when the binding site on mRNA form a strong secondary structure regardless of highly stable complementarity between ASO and the target.

J-GLOBAL

researchmap

Language：Japanese   Publishing type：Research paper, summary (national, other academic conference)  

In the drug discovery, it is necessary to explore a large compound database composed of several to hundred millions of compounds, thus the acceleration of docking calculation is greatly demanded. Reuse of calculation results is one of the feasible ways to accelerate. In this study, we focused on the fact that many of the compounds have common substructures, called fragments, and developed a fast docking tool specialized for the evaluation of a large number of compounds by reusing the docking calculation results of fragments. We propose a docking method to evaluate compounds efficiently by the calculation results of fragments, and we confirmed that the proposed docking method was approximately 8.4 times faster than AutoDock Vina keeping almost same accuracy.

J-GLOBAL

researchmap

Authorship：Lead author   Language：English  

CiNii Research

researchmap

Language：Japanese   Publishing type：Research paper, summary (national, other academic conference)  

In drug discovery research, it is necessary to explore a large compound database composed of several millions of compounds, thus acceleration of docking calculation is greatly demanded. Reuse of calculation results is one of feasible ways to accelerate. In this report, we focused on the fact that many of the compounds have common substructures, and developed a high speed docking tool by reusing the calculation results of substructures. By reordering the order of evaluation of compounds by a heuristic method and optimizing memory strategy by solving offline cache problem at high speed, it is possible to improve the reuse efficiency of the calculation results and the calculation speed was increased about 1.8 times faster.

J-GLOBAL

researchmap

Authorship：Lead author   Language：Japanese   Publishing type：Research paper, summary (national, other academic conference)  

The need to accelerate large - scale protein - ligand docking in virtual screening against a huge compound database led researchers to propose a strategy that entails memorizing the evaluation result of the partial structure of a compound and reusing it to evaluate other compounds. However, the previous method required frequent disk accesses, resulting in insufficient acceleration. Thus, more efficient memory usage can be expected to lead to further acceleration, and optimal memory usage could be achieved by solving the weighted offline cache problem. In this research, we propose an exact algorithm for the weighted offline cache problem, which we reduce to the minimum cost flow problem, and utilize the characteristics of the graph generated for this problem as constraints. The proposed algorithm was shown to be approximately seven times faster compared to an existing exact algorithm specified for directed acyclic graphs.

CiNii Books

CiNii Research

J-GLOBAL

researchmap

Authorship：Lead author   Language：Japanese   Publishing type：Research paper, summary (national, other academic conference)  

Protein-ligand docking is widely used as a computational method for drug candidate compound selection, however, the docking simulation is computationally expensive. Thus methods that filter out obviously unnecessary compounds prior to computationally expensive docking simulations have been proposed. In order to reduce the number of candidates, applying physicochemical thresholds like Lipinski's rule of five or filtering with machine learning are widely used. However, they are not considered to be combined with docking simulations. In this study, we proposed two methods presupposed to combine with docking simulation : (1) estimation of the upper limit of compound volume using a protein pocket size estimation method, and (2) estimation of the upper limit of compound volume by experimentally docking smaller amount of sample compounds. The sample compound docking method omitted 15% compounds overlooking only 0.5% dockable compounds.

J-GLOBAL

researchmap

Authorship：Lead author   Language：Japanese   Publishing type：Research paper, summary (national, other academic conference)  

Recently, the number of available protein tertiary structures and compounds has increased. However, structure-based virtual screening is computationally expensive due to docking simulations. Thus, methods that filter out obviously unnecessary compounds prior to computationally expensive docking simulations have been proposed. However, the calculation speed of these methods is not fast enough to evaluate more than 10 million compounds. In this study, we proposed a novel, docking-based pre-screening protocol named ESPRESSO (Extremely Speedy PRE-Screening method with Segmented cOmpounds). Partial structures (fragments) are often common among several compounds; therefore, the number of fragment variations needed for evaluation is smaller than that of compounds. Our method increased calculation speeds approximately 200-fold compared to conventional methods.

CiNii Books

CiNii Research

J-GLOBAL

researchmap

Language：Japanese   Publishing type：Research paper, summary (national, other academic conference)  

Virtual Screening (VS) is widely used in the process of a new drug development. A ligand-based method which is widely used in VS uses molecular descriptors of compounds of which inhibition activity for a target protein is proved. In ligand-based methods, many methods of machine learning or data mining are used. In this research, we constructed prediction models for target inhibition compounds by SVM and Deep Learning. In addition to the models, we constructed a prediction model by Deep Learning whose output layer is SVM. Finally, we applied these three models to prediction of human c-Yes kinase inhibitors and evaluated their accuracy.

CiNii Books

CiNii Research

J-GLOBAL

researchmap

Language：Japanese   Publishing type：Research paper, summary (national, other academic conference)   Publisher：Information Processing Society of Japan (IPSJ)  

Nowadays, drug development requires too much time and budget, and it is necessary to reduce them. In order to accept a compound as a new drug, it must be confirmed that it is metabolized and excreted. In this respect, one of the computational methods used for selecting compounds is drug clearance pathway prediction. This prediction method uses well-known drug's clearance pathway data as a training set. However data is expensive to get, and thus there are too few data. For this reason, we evaluated the usefulness of semi-supervised learning in this prediction problem, and tried to improve accuracy of this clearance pathway prediction. We also tried to add some features of compounds which are selected from 802 features by greedy algorithm to improve accuracy and evaluated their effect.

CiNii Books

CiNii Research

researchmap

J-GLOBAL

researchmap

Event date： 2025.10

Language：Japanese   Presentation type：Oral presentation (invited, special)  

Country：Japan  

Presentation ID: LS07

researchmap

Event date： 2025.10

Language：Japanese   Presentation type：Poster presentation  

Country：Japan  

PosterID: P01-26

researchmap

Event date： 2025.10

Language：Japanese   Presentation type：Poster presentation  

Country：Japan  

PosterID: P06-13

researchmap

Event date： 2025.10

Language：Japanese   Presentation type：Poster presentation  

Country：Japan  

PosterID: P06-15

researchmap

Event date： 2025.10

Language：Japanese   Presentation type：Poster presentation  

Country：Japan  

PosterID: P06-16

researchmap

Event date： 2025.10

Language：Japanese   Presentation type：Poster presentation  

Country：Japan  

PosterID: P06-19

researchmap

Event date： 2025.10

Language：Japanese   Presentation type：Poster presentation  

Country：Japan  

PosterID: P01-17

researchmap

Event date： 2025.10

Language：Japanese   Presentation type：Poster presentation  

Country：Japan  

PosterID: P01-06

researchmap

Event date： 2025.9

Language：Japanese   Presentation type：Poster presentation  

Country：Japan  

PosterID: 2Pos175

researchmap

Event date： 2025.9

Language：English   Presentation type：Oral presentation (invited, special)  

Venue：Hangzhou, Zhejiang   Country：China  

researchmap

Event date： 2025.9

Language：English   Presentation type：Poster presentation  

Venue：Hangzhou, Zhejiang   Country：China  

PosterID: P23

researchmap

Event date： 2025.9

Language：English   Presentation type：Poster presentation  

Venue：Hangzhou, Zhejiang   Country：China  

PosterID: P10

researchmap

Event date： 2025.9

Language：English   Presentation type：Poster presentation  

Venue：Hangzhou, Zhejiang   Country：China  

PosterID: P50

researchmap

Event date： 2025.9

Language：English   Presentation type：Poster presentation  

Venue：Hangzhou, Zhejiang   Country：China  

PosterID: P08

researchmap

Event date： 2025.9

Language：English   Presentation type：Poster presentation  

Venue：Hangzhou, Zhejiang   Country：China  

PosterID: P27

researchmap

Event date： 2025.6

Language：Japanese   Presentation type：Poster presentation  

Venue：Himeji, Hyogo   Country：Japan  

researchmap

Event date： 2025.6

Language：Japanese   Presentation type：Poster presentation  

Venue：Himeji, Hyogo   Country：Japan  

researchmap

Event date： 2025.6

Language：Japanese   Presentation type：Poster presentation  

Venue：Himeji, Hyogo   Country：Japan  

researchmap

Event date： 2025.6

Language：Japanese   Presentation type：Poster presentation  

Venue：Himeji, Hyogo   Country：Japan  

researchmap

Event date： 2025.6

Language：Japanese   Presentation type：Poster presentation  

Venue：Himeji, Hyogo   Country：Japan  

researchmap

Event date： 2024.12

Language：Japanese   Presentation type：Poster presentation  

Country：Japan  

researchmap

Event date： 2024.10

Language：Japanese   Presentation type：Poster presentation  

Country：Japan  

researchmap

Event date： 2024.10

Language：Japanese   Presentation type：Poster presentation  

Country：Japan  

researchmap

Event date： 2024.10

Language：Japanese   Presentation type：Poster presentation  

Country：Japan  

researchmap

Event date： 2024.10

Language：Japanese   Presentation type：Poster presentation  

Country：Japan  

researchmap

Event date： 2024.10

Language：Japanese   Presentation type：Poster presentation  

Country：Japan  

researchmap

Event date： 2024.10

Language：Japanese   Presentation type：Poster presentation  

Country：Japan  

researchmap

Event date： 2024.10

Language：Japanese   Presentation type：Poster presentation  

Country：Japan  

researchmap

Event date： 2024.10

Language：Japanese   Presentation type：Poster presentation  

Country：Japan  

researchmap

Event date： 2024.10

Language：English   Presentation type：Poster presentation  

Venue：NAHArt (Naha Cultural Arts Theater), Naha, Okinawa   Country：Japan  

researchmap

Event date： 2024.10

Language：English   Presentation type：Poster presentation  

Venue：NAHArt (Naha Cultural Arts Theater), Naha, Okinawa   Country：Japan  

researchmap

Event date： 2024.6

Language：English   Presentation type：Poster presentation  

Venue：Kyoto International Conference Center   Country：Japan  

Presentation ID: 28P-189

researchmap

Event date： 2024.6

Language：English   Presentation type：Poster presentation  

Venue：Kyoto International Conference Center   Country：Japan  

Presentation ID: 26P-210

researchmap

Event date： 2024.6

Language：Japanese   Presentation type：Oral presentation (invited, special)  

researchmap

Event date： 2024.6

Language：Japanese  

researchmap

Event date： 2024.6

Language：Japanese   Presentation type：Oral presentation (invited, special)  

researchmap

Event date： 2023.11

Language：Japanese   Presentation type：Poster presentation  

Country：Japan  

researchmap

Event date： 2023.11

Language：Japanese   Presentation type：Poster presentation  

Country：Japan  

researchmap

Event date： 2023.10

Language：Japanese   Presentation type：Poster presentation  

Country：Japan  

researchmap

Event date： 2023.9

Language：Japanese   Presentation type：Oral presentation (invited, special)  

Venue：柏の葉カンファレンスセンター   Country：Japan  

発表番号: WS-2

researchmap

Event date： 2022.10

Language：Japanese   Presentation type：Poster presentation  

Country：Japan  

researchmap

Event date： 2022.10

Language：Japanese   Presentation type：Oral presentation (general)  

Country：Japan  

Presentation ID: O2-1

researchmap

Event date： 2022.10

Language：Japanese   Presentation type：Oral presentation (general)  

Country：Japan  

Presentation ID: O3-2

researchmap

Event date： 2022.9

Language：Japanese   Presentation type：Poster presentation  

Country：Japan  

researchmap

Event date： 2022.6

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：つくば国際会議場   Country：Japan  

発表番号: O7-12

researchmap

Event date： 2021.11

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：仙台国際センター (オンライン開催)   Country：Japan  

発表番号: 2-03-1712

researchmap

Event date： 2021.10

Language：Japanese   Presentation type：Poster presentation  

researchmap

Event date： 2021.10

Language：Japanese   Presentation type：Oral presentation (general)  

Country：Japan  

Presentation ID: O2-1

researchmap

Event date： 2020.12

Language：Japanese   Presentation type：Oral presentation (invited, special)  

Venue：オンライン開催   Country：Japan  

発表番号: 2F-11

researchmap

Language：English   Presentation type：Oral presentation (invited, special)  

researchmap

Event date： 2019.10

Language：Japanese   Presentation type：Poster presentation  

Country：Japan  

researchmap

Event date： 2019.9

Language：Japanese   Presentation type：Poster presentation  

Country：Japan  

researchmap

Event date： 2019.2

Language：English  

researchmap

Event date： 2018.10

Language：Japanese   Presentation type：Poster presentation  

Country：Japan  

researchmap

Event date： 2018.9

Language：Japanese   Presentation type：Poster presentation  

Country：Japan  

researchmap

Event date： 2018.5

Language：Japanese   Presentation type：Poster presentation  

Country：Japan  

researchmap

Event date： 2018.2

Language：English   Presentation type：Poster presentation  

Venue：Moscone Center, San Francisco, CA   Country：United States  

researchmap

Event date： 2018.1

Language：English   Presentation type：Poster presentation  

Venue：Yokohama   Country：Japan  

Presentation ID: Poster C5

researchmap

Event date： 2017.9

Language：Japanese   Presentation type：Poster presentation  

Country：Japan  

researchmap

Event date： 2017.9

Language：Japanese   Presentation type：Poster presentation  

Country：Japan  

researchmap

Event date： 2016.10

Language：Japanese   Presentation type：Poster presentation  

Country：Japan  

researchmap

Event date： 2016.9 - 2016.10

Language：Japanese   Presentation type：Poster presentation  

Country：Japan  

researchmap

Event date： 2015.11

Language：Japanese   Presentation type：Poster presentation  

Country：Japan  

researchmap

Event date： 2015.11

Language：English   Presentation type：Poster presentation  

Country：Japan  

Presentation ID: P34

researchmap

Event date： 2015.10

Language：Japanese   Presentation type：Poster presentation  

Country：Japan  

researchmap

Applicant：国立大学法人東京工業大学

Application no：特願2021-031234  Date applied：2021.2

Patent/Registration no：特許第7057003号  Date registered：2022.4 

J-GLOBAL

researchmap

Applicant：国立大学法人東京工業大学

Application no：特願2021-023750  Date applied：2021.2

Announcement no：特開2022-078924  Date announced：2022.5

J-GLOBAL

researchmap

Applicant：アヘッド・バイオコンピューティング株式会社

Application no：特願2021-023750  Date applied：2021.2

Announcement no：特開2022-137148  Date announced：2022.9

Patent/Registration no：特許第7626466号  Date registered：2025.1 

J-GLOBAL

researchmap

Applicant：アヘッド・バイオコンピューティング株式会社

Application no：特願2021-023750  Date applied：2021.2

Announcement no：特開2022-078924  Date announced：2022.5

Patent/Registration no：特許第7125575号  Date registered：2022.8 

J-GLOBAL

researchmap

Applicant：東京工業大学

Application no：特願2020-189856  Date applied：2020.11

researchmap

researchmap

researchmap

researchmap

researchmap