Showing 1–14 of 14 results for author: Jiang, D

Token-Mol 1.0: Tokenized drug design with large language model

Authors: Jike Wang, Rui Qin, Mingyang Wang, Meijing Fang, Yangyang Zhang, Yuchen Zhu, Qun Su, Qiaolin Gou, Chao Shen, Odin Zhang, Zhenxing Wu, Dejun Jiang, Xujun Zhang, Huifeng Zhao, Xiaozhe Wan, Zhourui Wu, Liwei Liu, Yu Kang, Chang-Yu Hsieh, Tingjun Hou

Abstract: Significant interests have recently risen in leveraging sequence-based large language models (LLMs) for drug design. However, most current applications of LLMs in drug discovery lack the ability to comprehend three-dimensional (3D) structures, thereby limiting their effectiveness in tasks that explicitly involve molecular conformations. In this study, we introduced Token-Mol, a token-only 3D drug… ▽ More Significant interests have recently risen in leveraging sequence-based large language models (LLMs) for drug design. However, most current applications of LLMs in drug discovery lack the ability to comprehend three-dimensional (3D) structures, thereby limiting their effectiveness in tasks that explicitly involve molecular conformations. In this study, we introduced Token-Mol, a token-only 3D drug design model. This model encodes all molecular information, including 2D and 3D structures, as well as molecular property data, into tokens, which transforms classification and regression tasks in drug discovery into probabilistic prediction problems, thereby enabling learning through a unified paradigm. Token-Mol is built on the transformer decoder architecture and trained using random causal masking techniques. Additionally, we proposed the Gaussian cross-entropy (GCE) loss function to overcome the challenges in regression tasks, significantly enhancing the capacity of LLMs to learn continuous numerical values. Through a combination of fine-tuning and reinforcement learning (RL), Token-Mol achieves performance comparable to or surpassing existing task-specific methods across various downstream tasks, including pocket-based molecular generation, conformation generation, and molecular property prediction. Compared to existing molecular pre-trained models, Token-Mol exhibits superior proficiency in handling a wider range of downstream tasks essential for drug design. Notably, our approach improves regression task accuracy by approximately 30% compared to similar token-only methods. Token-Mol overcomes the precision limitations of token-only models and has the potential to integrate seamlessly with general models such as ChatGPT, paving the way for the development of a universal artificial intelligence drug design model that facilitates rapid and high-quality drug design by experts. △ Less

Submitted 19 August, 2024; v1 submitted 10 July, 2024; originally announced July 2024.

PPFlow: Target-aware Peptide Design with Torsional Flow Matching

Authors: Haitao Lin, Odin Zhang, Huifeng Zhao, Dejun Jiang, Lirong Wu, Zicheng Liu, Yufei Huang, Stan Z. Li

Abstract: Therapeutic peptides have proven to have great pharmaceutical value and potential in recent decades. However, methods of AI-assisted peptide drug discovery are not fully explored. To fill the gap, we propose a target-aware peptide design method called \textsc{PPFlow}, based on conditional flow matching on torus manifolds, to model the internal geometries of torsion angles for the peptide structure… ▽ More Therapeutic peptides have proven to have great pharmaceutical value and potential in recent decades. However, methods of AI-assisted peptide drug discovery are not fully explored. To fill the gap, we propose a target-aware peptide design method called \textsc{PPFlow}, based on conditional flow matching on torus manifolds, to model the internal geometries of torsion angles for the peptide structure design. Besides, we establish a protein-peptide binding dataset named PPBench2024 to fill the void of massive data for the task of structure-based peptide drug design and to allow the training of deep learning methods. Extensive experiments show that PPFlow reaches state-of-the-art performance in tasks of peptide drug generation and optimization in comparison with baseline models, and can be generalized to other tasks including docking and side-chain packing. △ Less

Submitted 16 June, 2024; v1 submitted 5 March, 2024; originally announced May 2024.

Comments: 18 pages

Deep Lead Optimization: Leveraging Generative AI for Structural Modification

Authors: Odin Zhang, Haitao Lin, Hui Zhang, Huifeng Zhao, Yufei Huang, Yuansheng Huang, Dejun Jiang, Chang-yu Hsieh, Peichen Pan, Tingjun Hou

Abstract: The idea of using deep-learning-based molecular generation to accelerate discovery of drug candidates has attracted extraordinary attention, and many deep generative models have been developed for automated drug design, termed molecular generation. In general, molecular generation encompasses two main strategies: de novo design, which generates novel molecular structures from scratch, and lead opt… ▽ More The idea of using deep-learning-based molecular generation to accelerate discovery of drug candidates has attracted extraordinary attention, and many deep generative models have been developed for automated drug design, termed molecular generation. In general, molecular generation encompasses two main strategies: de novo design, which generates novel molecular structures from scratch, and lead optimization, which refines existing molecules into drug candidates. Among them, lead optimization plays an important role in real-world drug design. For example, it can enable the development of me-better drugs that are chemically distinct yet more effective than the original drugs. It can also facilitate fragment-based drug design, transforming virtual-screened small ligands with low affinity into first-in-class medicines. Despite its importance, automated lead optimization remains underexplored compared to the well-established de novo generative models, due to its reliance on complex biological and chemical knowledge. To bridge this gap, we conduct a systematic review of traditional computational methods for lead optimization, organizing these strategies into four principal sub-tasks with defined inputs and outputs. This review delves into the basic concepts, goals, conventional CADD techniques, and recent advancements in AIDD. Additionally, we introduce a unified perspective based on constrained subgraph generation to harmonize the methodologies of de novo design and lead optimization. Through this lens, de novo design can incorporate strategies from lead optimization to address the challenge of generating hard-to-synthesize molecules; inversely, lead optimization can benefit from the innovations in de novo design by approaching it as a task of generating molecules conditioned on certain substructures. △ Less

Submitted 29 April, 2024; originally announced April 2024.

Influence of Material Parameter Variability on the Predicted Coronary Artery Biomechanical Environment via Uncertainty Quantification

Authors: Caleb C. Berggren, David Jiang, Y. F. Jack Wang, Jake A. Bergquist, Lindsay C. Rupp, Zexin Liu, Rob S. MacLeod, Akil Narayan, Lucas H. Timmins

Abstract: Central to the clinical adoption of patient-specific modeling strategies is demonstrating that simulation results are reliable and safe. Simulation frameworks must be robust to uncertainty in model input(s), and levels of confidence should accompany results. In this study we applied a coupled uncertainty quantification-finite element (FE) framework to understand the impact of uncertainty in vascul… ▽ More Central to the clinical adoption of patient-specific modeling strategies is demonstrating that simulation results are reliable and safe. Simulation frameworks must be robust to uncertainty in model input(s), and levels of confidence should accompany results. In this study we applied a coupled uncertainty quantification-finite element (FE) framework to understand the impact of uncertainty in vascular material properties on variability in predicted stresses. Univariate probability distributions were fit to material parameters derived from layer-specific mechanical behavior testing of human coronary tissue. Parameters were assumed to be probabilistically independent, allowing for efficient parameter ensemble sampling. In an idealized coronary artery geometry, a forward FE model for each parameter ensemble was created to predict tissue stresses under physiologic loading. An emulator was constructed within the UncertainSCI software using polynomial chaos techniques, and statistics and sensitivities were directly computed. Results demonstrated that material parameter uncertainty propagates to variability in predicted stresses across the vessel wall, with the largest dispersions in stress within the adventitial layer. Variability in stress was most sensitive to uncertainties in the anisotropic component of the strain energy function. Unary and binary interactions within the adventitial layer were the main contributors to stress variance, and the leading factor in stress variability was uncertainty in the stress-like material parameter summarizing contribution of the embedded fibers to the overall artery stiffness. Results from a patient-specific coronary model confirmed many of these findings. Collectively, this highlights the impact of material property variation on predicted artery stresses and presents a pipeline to explore and characterize uncertainty in computational biomechanics. △ Less

Submitted 26 January, 2024; originally announced January 2024.

Comments: To appear: Biomechanics and Modeling in Mechanobiology

Rounded notch method of femoral endarterectomy offers mechanical advantages in finite element models

Authors: David Jiang, Dongxu Liu, Efi Efrati, Nhung Nguyen, Luka Pocivavsek

Abstract: Objective: Use of a vascular punch to produce circular heel and toe arteriotomies for femoral endarterectomy with patch angioplasty is a novel technique. This study investigated the plausibility of this approach and the mechanical advantages of the technique using finite element models. Methods: The patient underwent a standard femoral endarterectomy. Prior to patch angioplasty, a 4.2 mm coronary… ▽ More Objective: Use of a vascular punch to produce circular heel and toe arteriotomies for femoral endarterectomy with patch angioplasty is a novel technique. This study investigated the plausibility of this approach and the mechanical advantages of the technique using finite element models. Methods: The patient underwent a standard femoral endarterectomy. Prior to patch angioplasty, a 4.2 mm coronary vascular punch was used to created proximal and distal circular arteriotomies. The idealized artery was modeled as a 9 mm cylinder with a central slit. The vertices of the slit were modeled as: a sharp V consistent with traditional linear arteriotomy, circular punched hole, and beveled punched hole. The artery was pressurized to achieve displacement consistent with the size of a common femoral artery prior to patch angioplasty. Maximum von Mises stress, area-averaged stress, and stress concentration factors were evaluated for all three models. Results: Maximum von Mises stress was 0.098 MPa with 5 mm of displacement and increased to 0.26 MPa with 10 mm of displacement. Maximum stress in the uniform circular model was 0.019 MPa and 0.018 with a beveled notch. Average stress was lowest in the circular punch model at 0.006 MP and highest in the linear V notch arteriotomy at 0.010 MPa. Stress concentration factor was significantly lower in both circular models compared with the V notch. Conclusions: Femoral endarterectomy modified with the creation of circular arteriotomies is a safe and effective surgical technique. Finite element modeling revealed reduced maximum von Mises stress and average stress at the vertices of a circular or beveled punch arteriotomy compared with a linear, V shaped arteriotomy. Reduced vertex stress may promote lower risk of restenosis. △ Less

Submitted 30 May, 2023; originally announced May 2023.

Infinite Physical Monkey: Do Deep Learning Methods Really Perform Better in Conformation Generation?

Authors: Haotian Zhang, Jintu Zhang, Huifeng Zhao, Dejun Jiang, Yafeng Deng

Abstract: Conformation Generation is a fundamental problem in drug discovery and cheminformatics. And organic molecule conformation generation, particularly in vacuum and protein pocket environments, is most relevant to drug design. Recently, with the development of geometric neural networks, the data-driven schemes have been successfully applied in this field, both for molecular conformation generation (in… ▽ More Conformation Generation is a fundamental problem in drug discovery and cheminformatics. And organic molecule conformation generation, particularly in vacuum and protein pocket environments, is most relevant to drug design. Recently, with the development of geometric neural networks, the data-driven schemes have been successfully applied in this field, both for molecular conformation generation (in vacuum) and binding pose generation (in protein pocket). The former beats the traditional ETKDG method, while the latter achieves similar accuracy compared with the widely used molecular docking software. Although these methods have shown promising results, some researchers have recently questioned whether deep learning (DL) methods perform better in molecular conformation generation via a parameter-free method. To our surprise, what they have designed is some kind analogous to the famous infinite monkey theorem, the monkeys that are even equipped with physics education. To discuss the feasibility of their proving, we constructed a real infinite stochastic monkey for molecular conformation generation, showing that even with a more stochastic sampler for geometry generation, the coverage of the benchmark QM-computed conformations are higher than those of most DL-based methods. By extending their physical monkey algorithm for binding pose prediction, we also discover that the successful docking rate also achieves near-best performance among existing DL-based docking models. Thus, though their conclusions are right, their proof process needs more concern. △ Less

Submitted 7 March, 2023; originally announced April 2023.

Relate auditory speech to EEG by shallow-deep attention-based network

Authors: Fan Cui, Liyong Guo, Lang He, Jiyao Liu, ErCheng Pei, Yujun Wang, Dongmei Jiang

Abstract: Electroencephalography (EEG) plays a vital role in detecting how brain responses to different stimulus. In this paper, we propose a novel Shallow-Deep Attention-based Network (SDANet) to classify the correct auditory stimulus evoking the EEG signal. It adopts the Attention-based Correlation Module (ACM) to discover the connection between auditory speech and EEG from global aspect, and the Shallow-… ▽ More Electroencephalography (EEG) plays a vital role in detecting how brain responses to different stimulus. In this paper, we propose a novel Shallow-Deep Attention-based Network (SDANet) to classify the correct auditory stimulus evoking the EEG signal. It adopts the Attention-based Correlation Module (ACM) to discover the connection between auditory speech and EEG from global aspect, and the Shallow-Deep Similarity Classification Module (SDSCM) to decide the classification result via the embeddings learned from the shallow and deep layers. Moreover, various training strategies and data augmentation are used to boost the model robustness. Experiments are conducted on the dataset provided by Auditory EEG challenge (ICASSP Signal Processing Grand Challenge 2023). Results show that the proposed model has a significant gain over the baseline on the match-mismatch track. △ Less

Submitted 20 March, 2023; originally announced March 2023.

All-Fibre Label-Free Nano-Sensor for Real-Time in situ Early Monitoring of Cellular Apoptosis

Authors: Danran Li, Nina Wang, Tianyang Zhang, Guangxing Wu, Yifeng Xiong, Qianqian Du, Yunfei Tian, Wei-wei Zhao, Jiandong Ye, Shulin Gu, Yanqing Lu, Dechen Jiang, Fei Xu

Abstract: The achievement of all-fibre functional nano-modules for subcellular label-free measurement has long been pursued due to the limitations of manufacturing techniques. In this paper, a compact all-fibre label-free nano-sensor composed of a fibre taper and zinc oxide nano-gratings is designed and applied for the early monitoring of apoptosis in single living cells. Because of its nanoscale dimensions… ▽ More The achievement of all-fibre functional nano-modules for subcellular label-free measurement has long been pursued due to the limitations of manufacturing techniques. In this paper, a compact all-fibre label-free nano-sensor composed of a fibre taper and zinc oxide nano-gratings is designed and applied for the early monitoring of apoptosis in single living cells. Because of its nanoscale dimensions, mechanical flexibility and minimal cytotoxicity to cells, the sensing module can be loaded in cells for long-term in situ tracking with high sensitivity. A gradual increase in the nuclear refractive index during the apoptosis process is observed, revealing the increase in molecular density and the decrease in cell volume. The strategy used in this study not only contributes to the understanding of internal environmental variations during cellular apoptosis but also provides a new platform for non-fluorescent all-fibre devices to investigate cellular events and to promote new progress in fundamental cell biochemical engineering. △ Less

Submitted 29 May, 2021; originally announced May 2021.

Steady-state joint distribution for first-order stochastic reaction kinetics

Authors: Youming Li, Da-Quan Jiang, Chen Jia

Abstract: While the analytical solution for the marginal distribution of a stochastic chemical reaction network has been extensively studied, its joint distribution, i.e. the solution of a high-dimensional chemical master equation, has received much less attention. Here we develop a novel method of computing the exact joint distributions of a wide class of first-order stochastic reaction systems in steady-s… ▽ More While the analytical solution for the marginal distribution of a stochastic chemical reaction network has been extensively studied, its joint distribution, i.e. the solution of a high-dimensional chemical master equation, has received much less attention. Here we develop a novel method of computing the exact joint distributions of a wide class of first-order stochastic reaction systems in steady-state conditions. The effectiveness of our method is validated by applying it to four gene expression models of biological significance, including models with 2A peptides, nascent mRNA, gene regulation, translational bursting, and alternative splicing. △ Less

Submitted 13 November, 2021; v1 submitted 8 February, 2021; originally announced February 2021.

Comments: 33 pages, 6 figures

Journal ref: Phys. Rev. E 104, 024408 (2021)

Phenotypic Equilibrium as Probabilistic Convergence in Multi-phenotype Cell Population Dynamics

Authors: Da-Quan Jiang, Yue Wang, Da Zhou

Abstract: We consider the cell population dynamics with $n$ different phenotypes. Both the Markovian branching process model (stochastic model) and the ordinary differential equation (ODE) system model (deterministic model) are presented, and exploited to investigate the dynamics of the phenotypic proportions. We will prove that in both models, these proportions will tend to constants regardless of initial… ▽ More We consider the cell population dynamics with $n$ different phenotypes. Both the Markovian branching process model (stochastic model) and the ordinary differential equation (ODE) system model (deterministic model) are presented, and exploited to investigate the dynamics of the phenotypic proportions. We will prove that in both models, these proportions will tend to constants regardless of initial population states ("phenotypic equilibrium") under weak conditions, which explains the experimental phenomenon in Gupta et al.'s paper. We also prove that Gupta et al.'s explanation is the ODE model under a special assumption. As an application, we will give sufficient and necessary conditions under which the proportion of one phenotype tends to $0$ (die out) or $1$ (dominate). We also extend our results to non-Markovian cases. △ Less

Submitted 2 November, 2022; v1 submitted 21 October, 2014; originally announced October 2014.

MSC Class: 60J85 (Primary); 92D25; 34D05 (Secondary)

Journal ref: PLOS ONE 12(2) (2017), e0170916

An allosteric model of the inositol trisphosphate receptor with nonequilibrium binding

Authors: Chen Jia, Daquan Jiang, Minping Qian

Abstract: The inositol trisphosphate receptor (IPR) is a crucial ion channel that regulates the Ca$^{2+}$ influx from the endoplasmic reticulum (ER) to the cytoplasm. A thorough study of the IPR channel contributes to a better understanding of calcium oscillations and waves. It has long been observed that the IPR channel is a typical biological system which performs adaptation. However, recent advances on t… ▽ More The inositol trisphosphate receptor (IPR) is a crucial ion channel that regulates the Ca$^{2+}$ influx from the endoplasmic reticulum (ER) to the cytoplasm. A thorough study of the IPR channel contributes to a better understanding of calcium oscillations and waves. It has long been observed that the IPR channel is a typical biological system which performs adaptation. However, recent advances on the physical essence of adaptation show that adaptation systems with a negative feedback mechanism, such as the IPR channel, must break detailed balance and always operate out of equilibrium with energy dissipation. Almost all previous IPR models are equilibrium models assuming detailed balance and thus violate the physical essence of adaptation. In this article, we constructed a nonequilibrium allosteric model of single IPR channels based on the patch-clamp experimental data obtained from the IPR in the outer membranes of isolated nuclei of the \emph{Xenopus} oocyte. It turns out that our model reproduces the patch-clamp experimental data reasonably well and produces both the correct steady-state and dynamic properties of the channel. Particularly, our model successfully describes the complicated bimodal [Ca$^{2+}$] dependence of the mean open duration at high [IP$_3$], a steady-state behavior which fails to be correctly described in previous IPR models. Finally, we used the patch-clamp experimental data to validate that the IPR channel indeed breaks detailed balance and thus is a nonequilibrium system which consumes energy. △ Less

Submitted 4 July, 2014; v1 submitted 11 November, 2013; originally announced November 2013.

Comments: 23 pages, 5 figures

Journal ref: Physical Biology, 11(5):056001, 2014

Overshoot in biological systems modeled by Markov chains: a nonequilibrium dynamic phenomenon

Authors: Chen Jia, Minping Qian, Daquan Jiang

Abstract: A number of biological systems can be modeled by Markov chains. Recently, there has been an increasing concern about when biological systems modeled by Markov chains will perform a dynamic phenomenon called overshoot. In this article, we found that the steady-state behavior of the system will have a great effect on the occurrence of overshoot. We confirmed that overshoot in general cannot occur in… ▽ More A number of biological systems can be modeled by Markov chains. Recently, there has been an increasing concern about when biological systems modeled by Markov chains will perform a dynamic phenomenon called overshoot. In this article, we found that the steady-state behavior of the system will have a great effect on the occurrence of overshoot. We confirmed that overshoot in general cannot occur in systems which will finally approach an equilibrium steady state. We further classified overshoot into two types, named as simple overshoot and oscillating overshoot. We showed that except for extreme cases, oscillating overshoot will occur if the system is far from equilibrium. All these results clearly show that overshoot is a nonequilibrium dynamic phenomenon with energy consumption. In addition, the main result in this article is validated with real experimental data. △ Less

Submitted 24 May, 2014; v1 submitted 10 November, 2013; originally announced November 2013.

Comments: 15 pages, 3 figures

Journal ref: IET Systems Biology, 8(4):138-145, 2014

Modeling stochastic phenotype switching and bet-hedging in bacteria: stochastic nonlinear dynamics and critical state identification

Authors: Chen Jia, Min-Ping Qian, Yu Kang, Da-Quan Jiang

Abstract: Fluctuating environments pose tremendous challenges to bacterial populations. It is observed in numerous bacterial species that individual cells can stochastically switch among multiple phenotypes for the population to survive in rapidly changing environments. This kind of phenotypic heterogeneity with stochastic phenotype switching is generally understood to be an adaptive bet-hedging strategy. M… ▽ More Fluctuating environments pose tremendous challenges to bacterial populations. It is observed in numerous bacterial species that individual cells can stochastically switch among multiple phenotypes for the population to survive in rapidly changing environments. This kind of phenotypic heterogeneity with stochastic phenotype switching is generally understood to be an adaptive bet-hedging strategy. Mathematical models are essential to gain a deeper insight into the principle behind bet-hedging and the pattern behind experimental data. Traditional deterministic models cannot provide a correct description of stochastic phenotype switching and bet-hedging, and traditional Markov chain models at the cellular level fail to explain their underlying molecular mechanisms. In this paper, we propose a nonlinear stochastic model of multistable bacterial systems at the molecular level. It turns out that our model not only provides a clear description of stochastic phenotype switching and bet-hedging within isogenic bacterial populations, but also provides a deeper insight into the analysis of multidimensional experimental data. Moreover, we use some deep mathematical theories to show that our stochastic model and traditional Markov chain models are essentially consistent and reflect the dynamic behavior of the bacterial system at two different time scales. In addition, we provide a quantitative characterization of the critical state of multistable bacterial systems and develop an effective data-driven method to identify the critical state without resorting to specific mathematical models. △ Less

Submitted 17 January, 2015; v1 submitted 9 November, 2013; originally announced November 2013.

Comments: 21 pages, 4 figures

Journal ref: Quantitative Biology, 2(3):110-125, 2014

Kinetic behavior of the general modifier mechanism of Botts and Morales with non-equilibrium binding

Authors: Chen Jia, Xu-Feng Liu, Min-Ping Qian, Da-Quan Jiang, Yu-Ping Zhang

Abstract: In this paper, we perform a complete analysis of the kinetic behavior of the general modifier mechanism of Botts and Morales in both equilibrium steady states and non-equilibrium steady states (NESS). Enlightened by the non-equilibrium theory of Markov chains, we introduce the net flux into discussion and acquire an expression of product rate in NESS, which has clear biophysical significance. Up t… ▽ More In this paper, we perform a complete analysis of the kinetic behavior of the general modifier mechanism of Botts and Morales in both equilibrium steady states and non-equilibrium steady states (NESS). Enlightened by the non-equilibrium theory of Markov chains, we introduce the net flux into discussion and acquire an expression of product rate in NESS, which has clear biophysical significance. Up till now, it is a general belief that being an activator or an inhibitor is an intrinsic property of the modifier. However, we reveal that this traditional point of view is based on the equilibrium assumption. A modifier may no longer be an overall activator or inhibitor when the reaction system is not in equilibrium. Based on the regulation of enzyme activity by the modifier concentration, we classify the kinetic behavior of the modifier into three categories, which are named hyperbolic behavior, bell-shaped behavior, and switching behavior, respectively. We show that the switching phenomenon, in which a modifier may convert between an activator and an inhibitor when the modifier concentration varies, occurs only in NESS. Effects of drugs on the Pgp ATPase activity, where drugs may convert from activators to inhibitors with the increase of the drug concentration, are taken as a typical example to demonstrate the occurrence of the switching phenomenon. △ Less

Submitted 29 September, 2011; v1 submitted 25 August, 2010; originally announced August 2010.

Comments: 19 pages, 10 figures

Journal ref: Journal of Theoretical Biology, 296:13-20, 2012