arXiv subjects

Junqing Wang

Publications and source records attributed to Junqing Wang.

9 recordsLinked to original sources

arXiv2026

EasyNano: rapid epitope-targeted nanobody CDR design via differentiable distogram optimization with ESMFold2

Computational design of nanobodies that bind user-specified protein epitopes could transform therapeutic development, but current methods either rely on stochastic sampling requiring days of GPU computation or inverse folding approaches unable to target epitopes directly. Here we present EasyNano, a practical pipeline for rapid, epitope-targeted nanobody complementarity-determining region (CDR) design that operates in approximately 10-20 minutes on a high-end personal workstation. EasyNano optimizes CDR residue logits via gradient descent through the ESMFold2 pairwise distance distogram, using the lightweight ESMFold2-Fast model (721M) as a differentiable oracle guided by a composite loss including a dedicated epitope proximity term. A full ESMFold2 (1.3B) CA-coordinate structure prior prevents framework pose drift. The wild-type logit initialization bias emerges as a critical practical parameter controlling CDR mutability. Across six target-framework pairs spanning self-recovery and de novo design scenarios, EasyNano improves ipTM by up to +0.559 -- from 0.143 to 0.702 (Ty1/RBD) -- and achieves a 4.6-fold improvement (ipTM 0.117 to 0.538) on a manually docked AQP4-targeting framework, while preserving ipTM on already-strong binders. Random CDR baselines (n=30 per target) confirm statistical significance (5.7 sigma above random mean for Ty1). Multi-seed analysis reveals diverse local minima, underscoring the importance of replicate runs. Kabsch cross-validation against crystal structures confirms that designed CDRs preserve the framework pose basin. EasyNano demonstrates that ESMFold2-based differentiable optimization provides a fast, practical, and epitope-specific approach to nanobody CDR design.

q-bio.QM↗

arXiv2026

Interpretable enzyme function prediction via sparse autoencoder features of ESMC across the microbial protein universe

Microbial genomes and metagenomes contain millions of proteins whose enzymatic functions remain unknown, the enzyme dark matter. While deep learning has improved protein function prediction, most methods are black boxes relying on sequence or structural similarity, limiting discovery of novel catalytic activities. The ESMC-6B protein language model and its sparse autoencoder with a 16,384-dimensional codebook of interpretable biological concepts, each annotated by GPT-5, creates a new opportunity: using these features directly as semantic signatures for enzyme function. Here, we show that ESMC-SAE features enable accurate and interpretable enzyme commission (EC) number prediction without task-specific training or GPU-intensive computation. On a balanced benchmark of 4,868 microbial SwissProt enzymes across 161 EC3 subclasses, ESMC-SAE binary features achieve 78.9% top-1 and 88.5% top-5 accuracy, 37.6% higher than 3-mer baselines (57.3%). In leave-one-EC3-class-out evaluation simulating discovery of novel enzyme classes, SAE features recover the EC1 superclass in 47.7% of cases (3.3x random, 14.3%), versus 26.6% for sequence methods. Discriminative features correspond to mechanistically interpretable concepts: catalytic triad geometry for hydrolases, NAD(P)H-binding Rossmann folds for oxidoreductases, phosphate-binding P-loops for transferases. We also survey the ESM Atlas of 7.7 million clusters and identify 169,859 dark enzyme-like candidates across all major microbial phyla. Our results establish a paradigm for enzyme function discovery in microbial dark matter: interpretable by design, scalable without GPU clusters, and applicable to the billions of proteins in the ESM Atlas.

q-bio.QM↗

arXiv2026

Modeling Vehicle-Type-Specific Pedestrian Crash Avoidance Behavior in Safety-Critical Interactions Using Smooth-Mamba Deep Reinforcement Learning

As automated vehicles (AVs) increasingly share roadways with human-driven vehicles (HDVs), understanding how pedestrians respond to different vehicle types in safety-critical interactions is essential for the safe deployment of automated driving technologies. This study extracts safety-critical pedestrian-vehicle interactions from the Argoverse 2 dataset to capture real-world crash avoidance behaviors in encounters involving AVs and HDVs. To model vehicle-type-specific pedestrian crash avoidance behavior, we develop a Smooth-Mamba Deep Deterministic Policy Gradient framework, termed SMamba-DDPG, which integrates smooth action constraints with efficient temporal representation learning. To quantify pedestrian behavioral differences, the framework trains separate crash avoidance policies for pedestrian interactions with AVs and HDVs. Results show that SMamba-DDPG outperforms baseline reinforcement learning and supervised learning models in reproducing pedestrian crash avoidance behaviors. Reconstructed trajectories demonstrate strong behavioral realism, accurately reproducing crash avoidance kinematics in both AV and HDV scenarios. Reaction time analysis shows that the model captures human-like response delays and reveals that pedestrians respond more quickly to AVs than to HDVs. Counterfactual analysis further indicates that pedestrians adopt lower crossing speeds when interacting with AVs. Large-scale safety analysis of model-generated data revealed that pedestrian-AV interactions consistently yielded lower conflict rates and higher pedestrian yielding rates compared to pedestrian-HDV interactions. The findings highlight the importance of incorporating vehicle-type-specific pedestrian behavioral models for safer automated driving system design and more realistic traffic simulations in mixed-traffic environments.

cs.AI↗

arXiv2026

TurboESM: Ultra-Efficient 3-Bit KV Cache Quantization for Protein Language Models with Orthogonal Rotation and QJL Correction

The rapid scaling of Protein Language Models (PLMs) has unlocked unprecedented accuracy in protein structure prediction and design, but the quadratic memory growth of the Key-Value (KV) cache during inference remains a prohibitive barrier for single-GPU deployment and high-throughput generation. While 8-bit quantization is now standard, 3-bit quantization remains elusive due to severe numerical outliers in activations. This paper presents TurboESM, an adaptation of Google's TurboQuant to the PLM domain. We solve the fundamental incompatibility between Rotary Position Embeddings (RoPE) and orthogonal transformations by deriving a RoPE-first rotation pipeline. We introduce a head-wise SVD calibration method tailored to the amino acid activation manifold, a dual look-up table (LUT) strategy for asymmetric K/V distributions, and a 1-bit Quantized Johnson-Lindenstrauss (QJL) residual correction. All experiments are conducted on ESM-2 650M, where our implementation achieves a 7.1x memory reduction (330 MB to 47 MB) while maintaining cosine similarity > 0.96 in autoregressive decoding across diverse protein families, including short peptides, transmembrane helices, enzyme active site fragments, and intrinsically disordered regions. We further implement a Triton-based fused decode attention kernel that eliminates intermediate dequantization memory allocations, achieving a 1.96x speedup over the PyTorch two-step path for the KV fetch operation alone; however, TurboESM incurs a prefill overhead of 21-27 ms relative to the original model due to KV quantization and packing, making it most suitable for memory-bound scenarios rather than latency-critical short-sequence workloads. Analysis reveals that PLMs exhibit sharper outlier profiles than large language models (LLMs) due to amino acid vocabulary sparsity, and our method effectively addresses these distributions.

q-bio.QM↗

arXiv2026

AntibodyDesignBFN: High-Fidelity Fixed-Backbone Antibody Design via Discrete Bayesian Flow Networks

The computational design of antibodies with high specificity and affinity is a cornerstone of modern therapeutic development. While deep generative models have demonstrated potential, they often struggle to balance high-fidelity geometric conditioning with the discrete nature of amino acid sequences. In this work, we present AntibodyDesignBFN, a novel framework for fixed-backbone antibody design based on Discrete Bayesian Flow Networks (BFN). Unlike standard diffusion models, BFNs operate on a continuous probability simplex, enabling a fully differentiable generative process that seamlessly integrates geometric gradients. By combining a lightweight Geometric Transformer with Invariant Point Attention (IPA) and a resource-efficient training strategy, our model establishes a new state-of-the-art. Evaluations on a rigorous 2025 temporal test set (43 complexes) demonstrate that AntibodyDesignBFN achieves an unprecedented Amino Acid Recovery(AAR) of 67.8%, significantly outperforming leading graph-based baselines. Furthermore, the model is highly efficient, enabling millisecond-scale inference on consumer-grade hardware. AntibodyDesignBFN thus offers a powerful, accessible, and mathematically robust framework for next generation antibody engineering. Code and model checkpoints are available at https://github.com/YueHuLab/AntibodyDesignBFN and https://huggingface.co/YueHuLab/AntibodyDesignBFN.

q-bio.QM↗

arXiv2025

Crypto-Assisted Graph Degree Sequence Release under Local Differential Privacy

Given a graph $G$ defined in a domain $\mathcal{G}$, we investigate locally differentially private mechanisms to release a degree sequence on $\mathcal{G}$ that accurately approximates the actual degree distribution. Existing solutions for this problem mostly use graph projection techniques based on edge deletion process, using a threshold parameter $θ$ to bound node degrees. However, this approach presents a fundamental trade-off in threshold parameter selection. While large $θ$ values introduce substantial noise in the released degree sequence, small $θ$ values result in more edges removed than necessary. Furthermore, $θ$ selection leads to an excessive communication cost. To remedy existing solutions' deficiencies, we present CADR-LDP, an efficient framework incorporating encryption techniques and differentially private mechanisms to release the degree sequence. In CADR-LDP, we first use the crypto-assisted Optimal-$θ$-Selection method to select the optimal parameter with a low communication cost. Then, we use the LPEA-LOW method to add some edges for each node with the edge addition process in local projection. LPEA-LOW prioritizes the projection with low-degree nodes, which can retain more edges for such nodes and reduce the projection error. Theoretical analysis shows that CADR-LDP satisfies $ε$-node local differential privacy. The experimental results on eight graph datasets show that our solution outperforms existing methods.

cs.CR↗

arXiv2024

Drone Data Analytics for Measuring Traffic Metrics at Intersections in High-Density Areas

This study employed over 100 hours of high-altitude drone video data from eight intersections in Hohhot to generate a unique and extensive dataset encompassing high-density urban road intersections in China. This research has enhanced the YOLOUAV model to enable precise target recognition on unmanned aerial vehicle (UAV) datasets. An automated calibration algorithm is presented to create a functional dataset in high-density traffic flows, which saves human and material resources. This algorithm can capture up to 200 vehicles per frame while accurately tracking over 1 million road users, including cars, buses, and trucks. Moreover, the dataset has recorded over 50,000 complete lane changes. It is the largest publicly available road user trajectories in high-density urban intersections. Furthermore, this paper updates speed and acceleration algorithms based on UAV elevation and implements a UAV offset correction algorithm. A case study demonstrates the usefulness of the proposed methods, showing essential parameters to evaluate intersections and traffic conditions in traffic engineering. The model can track more than 200 vehicles of different types simultaneously in highly dense traffic on an urban intersection in Hohhot, generating heatmaps based on spatial-temporal traffic flow data and locating traffic conflicts by conducting lane change analysis and surrogate measures. With the diverse data and high accuracy of results, this study aims to advance research and development of UAVs in transportation significantly. The High-Density Intersection Dataset is available for download at https://github.com/Qpu523/High-density-Intersection-Dataset.

eess.IV↗

arXiv2020

StreamNet: A DAG System with Streaming Graph Computing

To achieve high throughput in the POW based blockchain systems, researchers proposed a series of methods, and DAG is one of the most active and promising fields. We designed and implemented the StreamNet, aiming to engineer a scalable and endurable DAG system. When attaching a new block in the DAG, only two tips are selected. One is the parent tip whose definition is the same as in Conflux[1]; another is using Markov Chain Monte Carlo (MCMC) technique by which the definition is the same as IOTA [2]. We infer a pivotal chain along the path of each epoch in the graph, and a total order of the graph could be calculated without a centralized authority. To scale up, we leveraged the graph streaming property; high transaction validation speed will be achieved even if the DAG is growing. To scale out, we designed the direct signal gossip protocol to help disseminate block updates in the network, such that messages can be passed in the network more efficiently. We implemented our system based on IOTA's reference code (IRI) and ran comprehensive experiments over the different sizes of clusters of multiple network topologies.

cs.DC↗

arXiv2020

Knowledge-graph based Proactive Dialogue Generation with Improved Meta-Learning

Knowledge graph-based dialogue systems can narrow down knowledge candidates for generating informative and diverse responses with the use of prior information, e.g., triple attributes or graph paths. However, most current knowledge graph (KG) cover incomplete domain-specific knowledge. To overcome this drawback, we propose a knowledge graph based proactive dialogue generation model (KgDg) with three components, improved model-agnostic meta-learning algorithm (MAML), knowledge selection in knowledge triplets embedding, and knowledge aware proactive response generator. For knowledge triplets embedding and selection, we formulate it as a problem of sentence embedding to better capture semantic information. Our improved MAML algorithm is capable of learning general features from a limited number of knowledge graphs, which can also quickly adapt to dialogue generation with unseen knowledge triplets. Extensive experiments are conducted on a knowledge aware dialogue dataset (DuConv). The results show that KgDg adapts both fast and well to knowledge graph-based dialogue generation and outperforms state-of-the-art baseline.

cs.CL↗