In the development of targeted drugs, anticancer peptides (ACPs) have attracted great attention because of their high selectivity, low toxicity and minimal non-specificity. In this work, we report a framework of ACPs generation, which combines Wasserstein autoencoder (WAE) generative model and Particle Swarm Optimization (PSO) forward search algorithm guided by attribute predictive model to generate ACPs with desired properties. It is well known that generative models based on Variational AutoEncoder (VAE) and Generative Adversarial Networks (GAN) are difficult to be used for de novo design due to the problems of posterior collapse and difficult convergence of training. Our WAE-based generative model trains more successfully (lower perplexity and reconstruction loss) than both VAE and GAN-based generative models, and the semantic connections in the latent space of WAE accelerate the process of forward controlled generation of PSO, while VAE fails to capture this feature. Finally, we validated our pipeline on breast cancer targets (HIF-1) and lung cancer targets (VEGR, ErbB2), respectively. By peptide-protein docking, we found candidate compounds with the same binding sites as the peptides carried in the crystal structure but with higher binding affinity and novel structures, which may be potent antagonists that interfere with these target-mediated signaling.
This paper presents a comparative study of LLM-guided parameter tuning and Bayesian Optimization (BO) for automated PI parameter adjustment in Low Level Radio Frequency (LLRF) systems. Both methods demonstrate effective performance across continuous wave and pulsed operation modes, addressing the challenges of traditional manual tuning in superconducting accelerator facilities. Multi-cavity validation experiments were conducted on both operation modes. For continuous wave 2-parameter optimization, both methods converge to optimal regions within 15 iterations. For pulsed 6-parameter optimization involving coupled amplitude-phase control loops, both methods achieve satisfactory results in around 20 iterations, meeting stability requirements for beam operation. Comparative analysis reveals distinct application advantages: BO is better suited for large-scale batch optimization scenarios, particularly excelling in multi-cavity parallel commissioning and high-dimensional parameter space exploration due to its systematic probabilistic framework. The LLM-guided approach is more appropriate for online optimization scenarios with strict operational constraints, where its conservative optimization strategy and traceable reasoning process make it ideal for parameter fine-tuning during operations. The study provides practical method selection guidelines for intelligent operation of large-scale accelerator facilities.