rapid-silico-battery-electrolyte-electrochemical-reaction-generation

Rapid <em>in-silico</em> Battery Electrolyte Electrochemical Reaction Generation using 3T-VASP Multi-Scale Energy Minimization

Electrolytes are important components in lithium-ion batteries. However, battery degradation due to irreversible electrochemical reactions in the ...

Tencent Quantum Laboratory

Qubit initialization is a critical task in quantum computation and communication. Extensive efforts have been made to achieve this with high speed, efficiency and scalability. However, previous approaches have either been measurement-based and required fast feedback, suffered from crosstalk or required sophisticated calibration. Here, we report a fast and high-fidelity reset scheme, avoiding the issues above without any additional chip architecture. By modulating the flux through a transmon qubit, we realize a swap between the qubit and its readout resonator that suppresses the excited state population to 0.08% ± 0.08% within 34 ns (284 ns if photon depletion of the resonator is required). Furthermore, our approach (i) can achieve effective second excited state depletion, (ii) has negligible effects on neighboring qubits, and (iii) offers a way to entangle the qubit with an itinerant single photon, useful in quantum communication applications.

Rapid and unconditional parametric reset protocol for tunable superconducting qubits.

Quantum machine learning.

Shortcuts to adiabaticity are powerful quantum control methods, allowing quick evolution into target states of otherwise slow adiabatic dynamics. Such methods have widespread applications in quantum technologies, and various shortcuts to adiabaticity protocols have been demonstrated in closed systems. However, realizing shortcuts to adiabaticity for open quantum systems has presented a challenge due to the complex controls in existing proposals. Here, we present the experimental demonstration of shortcuts to adiabaticity for open quantum systems, using a superconducting circuit quantum electrodynamics system. By applying a counterdiabatic driving pulse, we reduce the adiabatic evolution time of a single lossy mode from 800 ns to 100 ns. In addition, we propose and implement an optimal control protocol to achieve fast and qubit-unconditional equilibrium of multiple lossy modes. Our results pave the way for precise time-domain control of open quantum systems and have potential applications in designing fast open-system protocols of physical and interdisciplinary interest, such as accelerating bioengineering and chemical reaction dynamics.

Shortcuts to adiabaticity for open systems in circuit quantum electrodynamics.

Drug resistance is a major threat to the global health and a significant concern throughout the clinical treatment of diseases and drug development. The mutation in proteins that is related to drug binding is a common cause for adaptive drug resistance. Therefore, quantitative estimations of how mutations would affect the interaction between a drug and the target protein would be of vital significance for the drug development and the clinical practice. Computational methods that rely on molecular dynamics simulations, Rosetta protocols, as well as machine learning methods have been proven to be capable of predicting ligand affinity changes upon protein mutation. However, the severely limited sample size and heavy noise induced overfitting and generalization issues have impeded wide adoption of machine learning for studying drug resistance. In this paper, we propose a robust machine learning method, termed SPLDExtraTrees, which can accurately predict ligand binding affinity changes upon protein mutation and identify resistance-causing mutations. Especially, the proposed method ranks training data following a specific scheme that starts with easy-to-learn samples and gradually incorporates harder and diverse samples into the training, and then iterates between sample weight recalculations and model updates. In addition, we calculate additional physics-based structural features to provide the machine learning model with the valuable domain knowledge on proteins for these data-limited predictive tasks. The experiments substantiate the capability of the proposed method for predicting kinase inhibitor resistance under three scenarios and achieve predictive accuracy comparable with that of molecular dynamics and Rosetta methods with much less computational costs.

SPLDExtraTrees: robust machine learning approach for predicting kinase inhibitor resistance.

The variational quantum eigensolver (VQE) is one of the most representative quantum algorithms in the noisy intermediate-scale quantum (NISQ) era, and is generally speculated to deliver one of the first quantum advantages for the ground-state simulations of some nontrivial Hamiltonians. However, short quantum coherence time and limited availability of quantum hardware resources in the NISQ hardware strongly restrain the capacity and expressiveness of VQEs. In this Letter, we introduce the variational quantum-neural hybrid eigensolver (VQNHE) in which the shallow-circuit quantum Ansatz can be further enhanced by classical post-processing with neural networks. We show that the VQNHE consistently and significantly outperforms the VQE in simulating ground-state energies of quantum spins and molecules given the same amount of quantum resources. More importantly, we demonstrate that, for arbitrary postprocessing neural functions, the VQNHE only incurs a polynomial overhead of processing time and represents the first scalable method to exponentially accelerate the VQE with nonunitary postprocessing that can be efficiently implemented in the NISQ era.

Variational Quantum-Neural Hybrid Eigensolver.

Lipophilicity (log) and aqueous solubility (log) play a central role in drug development. The accurate prediction of these properties remains to be solved due to data scarcity. Current methodologies neglect the intrinsic relationships between physicochemical properties and usually ignore the ionization effects. Here, we propose an attention-driven mixture-of-experts (MoE) model named ALipSol, which explicitly reproduces the hierarchy of task relationships. We adopt the principle of divide-and-conquer by breaking down the complex end point (log or log) into simpler ones (acidic p, basic p, and log) and allocating a specific expert network for each subproblem. Subsequently, we implement transfer learning to extract knowledge from related tasks, thus alleviating the dilemma of limited data. Additionally, we substitute the gating network with an attention mechanism to better capture the dynamic task relationships on a per-example basis. We adopt local fine-tuning and consensus prediction to further boost model performance. Extensive evaluation experiments verify the success of the ALipSol model, which achieves RMSE improvement of 8.04%, 2.49%, 8.57%, 12.8%, and 8.60% on the Lipop, ESOL, AqSolDB, external log, and external log data sets, respectively, compared with Attentive FP and the state-of-the-art  tools. In particular, our model yields more significant advantages (Welch's -test) for small training data, implying its high robustness and generalizability. The interpretability analysis proves that the atom contributions learned by ALipSol are more reasonable compared with the vanilla Attentive FP, and the substitution effects in benzene derivatives agreed well with empirical constants, revealing the potential of our model to extract useful patterns from data and provide guidance for lead optimization.

ALipSol: An Attention-Driven Mixture-of-Experts Model for Lipophilicity and Solubility Prediction.

Metalloproteins play indispensable roles in various biological processes ranging from reaction catalysis to free radical scavenging, and they are also pertinent to numerous pathologies including cancer, HIV infection, neurodegeneration, and inflammation. Discovery of high-affinity ligands for metalloproteins powers the treatment of these pathologies. Extensive efforts have been made to develop  approaches, such as molecular docking and machine learning (ML)-based models, for fast identification of ligands binding to heterogeneous proteins, but few of them have exclusively concentrated on metalloproteins. In this study, we first compiled the largest metalloprotein-ligand complex dataset containing 3079 high-quality structures, and systematically evaluated the scoring and docking powers of three competitive docking tools (, PLANTS, AutoDock Vina and Glide SP) for metalloproteins. Then, a structure-based deep graph model called MetalProGNet was developed to predict metalloprotein-ligand interactions. In the model, the coordination interactions between metal ions and protein atoms and the interactions between metal ions and ligand atoms were explicitly modelled through graph convolution. The binding features were then predicted by the informative molecular binding vector learned from a noncovalent atom-atom interaction network. The evaluation on the internal metalloprotein test set, the independent ChEMBL dataset towards 22 different metalloproteins and the virtual screening dataset indicated that MetalProGNet outperformed various baselines. Finally, a noncovalent atom-atom interaction masking technique was employed to interpret MetalProGNet, and the learned knowledge accords with our understanding of physics.

MetalProGNet: a structure-based deep graph model for metalloprotein-ligand interaction predictions.

Discrete time crystals (DTCs) have recently attracted increasing attention, but most DTC models and their properties are only revealed after disorder average. In this Letter, we propose a simple disorder-free periodically driven model that exhibits nontrivial DTC order stabilized by Stark many-body localization (MBL). We demonstrate the existence of the DTC phase by analytical analysis from perturbation theory and convincing numerical evidence from observable dynamics. The new DTC model paves a new promising way for further experiments and deepens our understanding of DTCs. Since the DTC order does not require special quantum state preparation and the strong disorder average, it can be naturally realized on the noisy intermediate-scale quantum hardware with much fewer resources and repetitions. Moreover, in addition to the robust subharmonic response, there are other novel robust beating oscillations in the Stark-MBL DTC phase that are absent in random or quasiperiodic MBL DTCs.

Discrete Time Crystal Enabled by Stark Many-Body Localization.

Mutation-induced drug resistance is a significant challenge to the clinical treatment of many diseases, as structural changes in proteins can diminish drug efficacy. Understanding how mutations affect protein-ligand binding affinities is crucial for developing new drugs and therapies. However, the lack of a large-scale and high-quality database has hindered the research progresses in this area. To address this issue, we have developed MdrDB, a database that integrates data from seven publicly available datasets, which is the largest database of its kind. By integrating information on drug sensitivity and cell line mutations from Genomics of Drug Sensitivity in Cancer and DepMap, MdrDB has substantially expanded the existing drug resistance data. MdrDB is comprised of 100,537 samples of 240 proteins (which encompass 5119 total PDB structures), 2503 mutations, and 440 drugs. Each sample brings together 3D structures of wild type and mutant protein-ligand complexes, binding affinity changes upon mutation (ΔΔG), and biochemical features. Experimental results with MdrDB demonstrate its effectiveness in significantly enhancing the performance of commonly used machine learning models when predicting ΔΔG in three standard benchmarking scenarios. In conclusion, MdrDB is a comprehensive database that can advance the understanding of mutation-induced drug resistance, and accelerate the discovery of novel chemicals.

A mutation-induced drug resistance database (MdrDB).

Acid-base dissociation constant (p) is a key physicochemical parameter in chemical science, especially in organic synthesis and drug discovery. Current methodologies for p prediction still suffer from limited applicability domain and lack of chemical insight. Here we present MF-SuP-p (multi-fidelity modeling with subgraph pooling for p prediction), a novel p prediction model that utilizes subgraph pooling, multi-fidelity learning and data augmentation. In our model, a knowledge-aware subgraph pooling strategy was designed to capture the local and global environments around the ionization sites for micro-p prediction. To overcome the scarcity of accurate p data, low-fidelity data (computational p) was used to fit the high-fidelity data (experimental p) through transfer learning. The final MF-SuP-p model was constructed by pre-training on the augmented ChEMBL data set and fine-tuning on the DataWarrior data set. Extensive evaluation on the DataWarrior data set and three benchmark data sets shows that MF-SuP-p achieves superior performances to the state-of-the-art p prediction models while requires much less high-fidelity training data. Compared with Attentive FP, MF-SuP-p achieves 23.83% and 20.12% improvement in terms of mean absolute error (MAE) on the acidic and basic sets, respectively.

MF-SuP-p: Multi-fidelity modeling with subgraph pooling mechanism for p prediction.

Protein-ligand binding affinity (PLBA) prediction is the fundamental task in drug discovery. Recently, various deep learning-based models predict binding affinity by incorporating the three-dimensional (3D) structure of protein-ligand complexes as input and achieving astounding progress. However, due to the scarcity of high-quality training data, the generalization ability of current models is still limited. Although there is a vast amount of affinity data available in large-scale databases such as ChEMBL, issues such as inconsistent affinity measurement labels (i.e. IC50, Ki, Kd), different experimental conditions, and the lack of available 3D binding structures complicate the development of high-precision affinity prediction models using these data. To address these issues, we (i) propose Multi-task Bioassay Pre-training (MBP), a pre-training framework for structure-based PLBA prediction; (ii) construct a pre-training dataset called ChEMBL-Dock with more than 300k experimentally measured affinity labels and about 2.8M docked 3D structures. By introducing multi-task pre-training to treat the prediction of different affinity labels as different tasks and classifying relative rankings between samples from the same bioassay, MBP learns robust and transferrable structural knowledge from our new ChEMBL-Dock dataset with varied and noisy labels. Experiments substantiate the capability of MBP on the structure-based PLBA prediction task. To the best of our knowledge, MBP is the first affinity pre-training model and shows great potential for future development. MBP web-server is now available for free at: https://huggingface.co/spaces/jiaxianustc/mbp.

Multi-task bioassay pre-training for protein-ligand binding affinity prediction.

Quantum computing, with its superior computational capabilities compared to classical approaches, holds the potential to revolutionize numerous scientific domains, including pharmaceuticals. However, the application of quantum computing for drug discovery has primarily been limited to proof-of-concept studies, which often fail to capture the intricacies of real-world drug development challenges. In this study, we diverge from conventional investigations by developing a hybrid quantum computing pipeline tailored to address genuine drug design problems. Our approach underscores the application of quantum computation in drug discovery and propels it towards more scalable system. We specifically construct our versatile quantum computing pipeline to address two critical tasks in drug discovery: the precise determination of Gibbs free energy profiles for prodrug activation involving covalent bond cleavage, and the accurate simulation of covalent bond interactions. This work serves as a pioneering effort in benchmarking quantum computing against veritable scenarios encountered in drug design, especially the covalent bonding issue present in both of the case studies, thereby transitioning from theoretical models to tangible applications. Our results demonstrate the potential of a quantum computing pipeline for integration into real world drug design workflows.

A hybrid quantum computing pipeline for real world drug discovery.

Ab-initio methods such as density functional theory (DFT) is useful for fundamental atomistic-level study and is widely used across many scientific fields, including for the discovery of electrochemical reaction byproducts. However, many DFT steps may be needed to discover rare electrochemical reaction byproducts, which limits DFT's scalability. In this work, we demonstrate that it is possible to generate many elementary electrochemical reaction byproducts in-silico using just a small number of ab-initio energy minimization steps if it is done in a multi-scale manner, such as via previously reported tiered tensor transform (3T) method. We first demonstrate the algorithm through a simple example of a complex floppy organic molecule passivator binding onto perovskite solar cell surface defect site. We then demonstrate more complex examples by generating hundreds of electrochemical reaction byproducts in lithium-ion battery liquid electrolyte (many are verified in previous experimental studies), with most trajectories completed within 50-100 DFT steps as opposed to more than 10,000 steps typically utilized in an ab-initio molecular dynamics trajectory. This approach requires no machine learning training data generation and can be directly applied on any new chemistries, making it suitable for ab-initio elementary chemical reaction byproduct investigation when temperature dependence is not required.

Article	Year
Rapid and unconditional parametric reset protocol for tunable superconducting qubits. Nature communications\| PubMed ID: 34635663	2021
Quantum machine learning. National science review\| PubMed ID: 34691828	2019
Shortcuts to adiabaticity for open systems in circuit quantum electrodynamics. Nature communications\| PubMed ID: 35013301	2022
SPLDExtraTrees: robust machine learning approach for predicting kinase inhibitor resistance. Briefings in bioinformatics\| PubMed ID: 35262669	2022
Variational Quantum-Neural Hybrid Eigensolver. Physical review letters\| PubMed ID: 35394326	2022
ALipSol: An Attention-Driven Mixture-of-Experts Model for Lipophilicity and Solubility Prediction. Journal of chemical information and modeling\| PubMed ID: 36417544	2022
MetalProGNet: a structure-based deep graph model for metalloprotein-ligand interaction predictions. Chemical science\| PubMed ID: 36845922	2023
Discrete Time Crystal Enabled by Stark Many-Body Localization. Physical review letters\| PubMed ID: 37027857	2023
A mutation-induced drug resistance database (MdrDB). Communications chemistry\| PubMed ID: 37316673	2023
MF-SuP-p: Multi-fidelity modeling with subgraph pooling mechanism for p prediction. Acta pharmaceutica Sinica. B\| PubMed ID: 37425064	2023
Multi-task bioassay pre-training for protein-ligand binding affinity prediction. Briefings in bioinformatics\| PubMed ID: 38084920	2023
A hybrid quantum computing pipeline for real world drug discovery. Scientific reports\| PubMed ID: 39043787	2024
3T-VASP: fast ab-initio electrochemical reactor via multi-scale gradient energy minimization. Nature communications\| PubMed ID: 39578465	2024

Shengyu Zhang

Biography

JoVE Journal Publications

Other Publications