Predicting cellular responses to combinatorial perturbations is a central challenge in drug discovery and therapeutic development. The combinatorial explosion of possible treatments renders exhaustive experimental screening infeasible, necessitating accurate predictive models. However, existing computational models often rely on oversimplified additiveeffect assumptions and struggle to distinguish perturbations that produce similar yet distinct effects within latent spaces. To overcome these limitations, we introduce scCAVAE, a deep learning framework that integrates a multi-head attention encoder to capture nonlinear synergistic and antagonistic interactions. The framework employs a novel hierarchical supervised contrastive learning strategy to engineer a semantically structured latent space by encoding intrinsic hierarchical relationships (e.g., inclusion, overlap, and disjointness) among perturbation combinations. Evaluation across multiple public benchmark datasets demonstrates that scCAVAE consistently outperforms existing methods, particularly in out-of-distribution (OOD) prediction tasks. Our findings provide a robust computational tool for high-throughput perturbation screening and establish a novel paradigm for modeling complex cellular responses. Our code is publicly available (https://github.com/cskyan/scCAVAE).

RNA-binding proteins (RBPs) play crucial roles in gene regulation. Their dysregulation has been increasingly linked to neurodegenerative diseases, liver cancer, and lung cancer. Although experimental methods like CLIP-seq accurately identify RNA-protein binding sites, they are time-consuming and costly. To address this, we propose RMDNet-a deep learning framework that integrates CNN, CNN-Transformer, and ResNet branches to capture features at multiple sequence scales. These features are fused with structural representations derived from RNA secondary structure graphs. The graphs are processed using a graph neural network with DiffPool. To optimize feature integration, we incorporate an improved dung beetle optimization algorithm, which adaptively assigns fusion weights during inference. Evaluations on the RBP-24 benchmark show that RMDNet outperforms state-of-the-art models including GraphProt, DeepRKE, and DeepDW across multiple metrics. On the RBP-31 dataset, it demonstrates strong generalization ability, while ablation studies on RBPsuite2.0 validate the contributions of individual modules. We assess biological interpretability by extracting candidate binding motifs from the first-layer CNN kernels. Several motifs closely match experimentally validated RBP motifs, confirming the model's capacity to learn biologically meaningful patterns. A downstream case study on YTHDF1 focuses on analyzing interpretable spatial binding patterns, using a large-scale prediction dataset and CLIP-seq peak alignment. The results confirm that the model captures localized binding signals and spatial consistency with experimental annotations. Overall, RMDNet is a robust and interpretable tool for predicting RNA-protein binding sites. It has broad potential in disease mechanism research and therapeutic target discovery. The source code is available https://github.com/cskyan/RMDNet .

One of the main roadblocks to Atomic Force Microscope (AFM) based nanomanipulation is lack of real time visual feedback. Although the model based visual feedback can partly solve this problem, its unguaranteed reliability due to the inaccurate models in nano-environment still limits the efficiency of AFM based nanomanipulation. This paper introduce a Real-time Fault Detection and Correction (RFDC) method to improve the reliability of the visual feedback. By utilizing Kalman filter and local scan technologies, the RFDC method not only can real-time detect the fault display caused by the modeling error, but also can on-line correct it without interrupting manipulation. In this way, the visual feedback keeps consistent with the true environment changes during manipulation, which makes several operations being finished without a image scanning in between. The theoretical study and the implementation of the RFDC method are elaborated in this paper. Experiments of manipulating nano-particles have been carried out to demonstrate the effectiveness and efficiency of the proposed method.

In this paper, we present a new mathematical model for designing micro electrode geometry and electric field for carbon nanotubes application. The micro electrode design is based on the assumption that the electrical potential at any point (x,y,z) created by a micro electrode of interest is defined by a polynomial that obeys Laplacepsilas equation. By substituting this polynomial into Laplacepsilas equation the corresponding equipotentials can therefore be determined, and these in turn can be used to define the required micro electrode boundaries for use in carbon nanotube deposition, manipulation, and implementation using dielectrophoresis for electronics and sensing application. Simulation and preliminary experimental results are presented for the developed models.

Atomic Force Microscope (AFM) has been used as a manipulation tool for a decade. The problem of lacking real time visual feedback still limits its efficiency and hinders its wide application. Although the model based visual feedback can partly solve this problem, due to the complexity of nano environment, it is difficult to use a model to accurately describe the object’s behavior. The modeling error will give the operator a false feedback and lead to a failed manipulation. In this paper, a strategy for visual feedback error on-line detection and correction is proposed to solve this problem. As the real time force information is a key factor for this strategy, an adaptable end effector is employed to accurately measure the interaction force between the probe and the nano-objects, and the system error is also compensated to improve the accuracy of interaction force measurement. Based on the true real time force information, an extended Kalman filter is developed to online detect whether there is a false feedback. Once a false feedback is detected, an optimal searching pattern is generated to get the real manipulation result in a short time. With the assistance of this strategy, the false visual feedback can be realtime detected and corrected without interrupting manipulation. Complex manipulation task can be finished without being interrupted by a new image scan. Experiments of manipulating nano-particles are performed to verify the effectiveness of this strategy, which demonstrated the improved efficiency of the AFM based nano-assembly system.

In this paper, a new pneumatic end effector system for micro/nano fluidic handling, nanomanufacturing, and micro/nano manipulation is presented. The new micro pneumatic end-effector system consists of a DC micro-diaphragm pump and compressor, one region of flexible latex tube, a Polyvinylidene Fluoride (PVDF) sensor for in-situ measurement of micro force, and a micro steel tip. The micro steel tip of the new pneumatic end effector system has an internal diameter (ID) of 20 mum used for handling nano entities such as carbon nanotubes, DNA, micro/nano particles as well as for microfluidic handling and droplet control. The DC micro-diaphragm pump is automatically controlled via a voltage driver interfaced with a computer in order to effectively and efficiently control suction force and pressure during microfluidic handling and droplet control in nano manufacturing. The new pneumatic end effector system with force sensing can significantly improve the success rate for handling/depositing micro/nano entities in the case of carbon nanotubes. The experimental results show the success rate of placing carbon nanotubes between electrodes can reach close to 80%. Ultimately, the technology will provide a critical and major step towards the development of automated manufacturing process for batch assembly of micro devices, manufacturing of nano devices, microfluidic droplet control, and drug delivery

An atomic force microscope (AFM) explores the topography of a sample surface using a micro-sized flexible cantilever, which works as a flexible robot arm. The flexible cantilever is controlled to keep vibrating when an AFM works in the tapping mode. The cantilever is modeled as a flexible beam instead of a point mass system in this paper. The nonlinear interaction force between the tip and sample surface is also modeled. A simulation environment is developed to analyze the dynamics of cantilevers using the flexible beam model. Simulation results confirm that the flexible beam model can represent the system more accurately than the point-mass model. It has been shown that lower modes are more sensitive to changes of surface topography or surface materials when the cantilever is driven to vibrate at a higher harmonic mode. At the same time, this simulation environment also provides a more accurate way to validate the design of a new AFM probe and AFM controller than simulation packages which use the point-mass model

Nanomanufacturing using Atomic Force Microcopy has been widely investigated. Most of nanomanipulation schemes go through the scan-design-manipulation-scan cycle manually which is time consuming and inefficient. Automated AFM tip path planning is desirable for nanomanufacturing, but does not receive much attention. In this paper, a CAD guided automated nanomanufacturing system is presented. Based on the CAD model of a nanostructure, the manipulation paths for both nanoparticles and nanorods are generated automatically. A local scanning method is developed to compensate for the random drift that may cause the failure of the nano-assembly. The experimental results demonstrate that the developed algorithm can be employed to manufacture nanostructures efficiently. The research work opens a door to the CAD guided automated nanomanufacturing.

Nanomanipulation using Atomic Force Microscope (AFM) has been extensively investigated for many years. However, control of tip position during nanomanipulation is still a major issue because of the deformation of the cantilever caused by manipulation force. The softness of the conventional cantilevers also cause the failure of the manipulation of relatively large and sticky nano-object because the tip can easily slip over the nano-object. In this paper, an active atomic force microscopy probe is used to solve these problems by changing the cantilever’s flexibility or rigidity through different control strategies in imaging and manipulation modes respectively. During imaging mode, the active probe is controlled to bend up with respect to the interaction force between the tip and samples, thus making the tip response faster and increase the imaging speed. During manipulation mode, the active probe is controlled to bend down with respect to the interaction force between tip and the samples; thus increasing its nominal rigidity to avoid tip slipping over object. A detailed model of the active probe is presented in this paper and the controller designed based on the proposed active probe model is also implemented on the augmented reality system, which is an AFM based nanomanipulation system with both real-time visual and haptic feedback. The simulation results for the control strategies and the preliminary experimental results for the AFM based nanmomanipulation verified the validity of the model and effectiveness of the controller.

Atomic force microscope (AFM) based nanomanipulation has been extensively investigated for many years. However, the efficiency and accuracy of the AFM based nanomanipulation is still a major issue due to the nonlinearities and uncertainties in nanomanipulation operations. The deformation of the cantilever caused by manipulation force, in our experience, is one of the most major nonlinearities and uncertainties. It causes difficulties in precisely controlling the tip position, which will cause the tip to miss the position of the object. In order to solve this problem, the traditional approach is to use a rigid cantilever. However, this will significantly reduce the sensitivity of the force feeling during the manipulation, which is essential for achieving an efficient and reliable nanomanipulation. An active AFM probe is used to solve this problem by directly controlling the cantilever's flexibility or rigidity during manipulation. An infinite dimensional model of the active probe is developed. Control of the active probe employing an optimal LQR control law is also implemented. The experimental results have verified the theoretical model and demonstrated that the precise position control and high sensitive interaction force measurement can be achieved simultaneously.