Package: PERCEPTIONx 0.3.0

Contents | 0. Requirements & Launch | 0.1 Requirements | 0.2 Launch | 0.3 Overall Flow | 1. Interface Overview | 2. Data Tab: Loading Data | 2.1 Loading DepMap Reference Data (required for training) | 2.2 Loading Models | 2.3 Uploading an Expression Matrix (patient scRNA-seq) | 2.4 Uploading the Cell-to-Patient Map (Mapping) | 2.5 Uploading Clinical Responses (Response, optional but recommended) | 2.6 Clustering (Seurat) | 3. Train Tab: Training Models (Optional) | 3.1 Parameters | 3.2 Interpreting the Results | 4. Predict Tab: Predicting Viability Scores | 5. Visualize Tab | 5.1 Clone Distribution (stacked bar chart) | 5.2 Clone Viability (lollipop plot) | 5.3 ROC Curve | 5.4 Response Boxplot (responders vs. non-responders) | 5.5 Model Performance | 5.6 Spatial Plots (UMAP / t-SNE) | 5.7 High-Resolution Download | 6. Help Tab | 7. FAQ | 8. Citation & Contact

目录 | 0. 环境要求与启动 | 0.1 环境要求 | 0.2 启动方式 | 0.3 整体流程 | 1. 界面总览 | 2. Data 模块：加载数据 | 2.1 加载 DepMap 参考数据（训练必需） | 2.2 加载模型 | 2.3 上传表达矩阵（患者 scRNA-seq） | 2.4 上传细胞-患者映射（Mapping） | 2.5 上传临床响应（Response，可选但推荐） | 2.6 聚类分群（Seurat） | 3. Train 模块：训练模型（可选） | 3.1 参数说明 | 3.2 结果解读 | 4. Predict 模块：预测杀伤值 | 5. Visualize 模块：可视化 | 5.1 Clone Distribution（stackplot） | 5.2 Clone Viability（lolliplot） | 5.3 ROC Curve | 5.4 Response Boxplot（R vs NR） | 5.5 Model Performance | 5.6 空间图（UMAP / t-SNE） | 5.7 高清下载 | 6. Help 页 | 7. 常见问题（FAQ） | 8. 引用与联系

1. Overview | 2. Minimal Workflow | 2.1 Load the DepMap reference data | 2.2 Load pre-trained models | 2.3 Prepare the clone-level expression data | 2.4 Predict drug sensitivity | 2.4.1 Clone-level prediction | 2.4.2 Patient-level aggregation | 2.5 Visualize | 3. Advanced Topics | 3.1 Train your own models | 3.2 Evaluate model performance | 3.2.1 Aggregation modes | 3.2.2 Pseudo-bulk expression | 3.3 Validate predictions on DepMap cell lines | 3.4 The plotting suite | 3.5 prepare_data() and Seurat clustering | 3.6 The Shiny web application | 3.7 End-to-end pipeline script | 3.8 Agent skills | 4. Citation