Datasets and Rscripts for : Multiomic and Longitudinal Dissection of Immune Dynamics Associated with Parkinsonism after Ciltacabtagene Autoleucel Therapy

Seurat Objects ds_indexContains all cells recovered via scRNA-seq and scTCR-seq from cerebrospinal fluid (CSF) and peripheral blood mononuclear cells (PBMCs) of the index patient who developed parkinsonism following CAR-T cell therapy. Samples were collected longitudinally across multiple time points (day 20 to day 204 post–CAR T). T_cells_indexSubset of ds_index comprising T cells identified from the combined CSF and PBMC samples of the index patient. ds_controlContains all cells recovered from peripheral blood (PBMCs) via scRNA-seq and scTCR-seq from: the index patient at early time points (pre–CAR T and day 7 post–CAR T), and three additional control patients treated with cilta-cel, sampled at matched time points (pre–CAR T and day 9/10 post–CAR T). T_cells_controlSubset of ds_control comprising T cells identified from PBMC samples of the index patient (pre–CAR T and day 7 post–CAR T) and the three control patients treated with cilta-cel (pre–CAR T and day 9/10 post–CAR T). R Scripts Additionally, three R scripts are provided: Script_index_QC_combination_of_index_datasets (Quality Control index patient)Details the quality control, normalization, integration, applied to the index patient datasets. Script_index_analysis_ds_T_cell_subset (Visualization index patient)Contains the code used to generate figures and perform downstream analyses of the index patient dataset, including clonal and phenotypic characterization. Script_control_QC_and_downstream (Early time point and control cohort processing script)Describes the quality control, Seurat object creation, and visualization steps for early time point samples (ds_control and T_cells_control) from the index patient and the control cohort. Methods Summary for Lentiviral Integration Analysis Data Preprocessing: Adapter Removal and Quality Trimming Raw sequencing reads were processed using cutadapt to remove adapter sequences. Further trimming for low-quality bases and length filtering was performed using Trimmomatic with the parameters SLIDINGWINDOW:4:20 and MINLEN:50. Filter Out CAR Construct Reads Reads containing the CAR construct sequence were identified and removed using grep and seqtk to ensure only genomic sequences of interest remained for further analysis. Mapping and Alignment Genome Reference Construction A custom genome reference (hg19_lenti.fa) was built by appending the lentiviral sequence to the human genome (hg19) reference. Alignment Trimmed reads were aligned to the custom reference genome using BWA-MEM version 0.7.17 for high sensitivity. SAMtools v1.21 was used to convert and sort alignment files for downstream processing. Identification of Chimeric Reads Chimeric reads (spanning human and lentiviral integration sites) were extracted using samtools and processed further. Integration Site Analysis Annotation of Integration Sites Significant integration sites were determined using BEDTools v2.31 to identify overlap with genomic features. Annotated integration sites were further processed using HOMER v5.1 to classify insertions into genomic regions (e.g., promoters, introns, exons, intergenic regions). Visualization and Results Chromosomal Distribution Integration site distributions were visualized across chromosomes, excluding haploid content using a python script. Feature Enrichment Enrichment analysis for genomic features (e.g., TEs, promoters) was performed using observed-to-expected ratios and log2 enrichment scores, using HOMER. Summary This workflow allowed us to identify approximately 19,046 unique lentiviral integration sites from a population of ~1 million cells. The analysis confirmed the preference of lentiviral vectors for active genomic regions, including transcriptional elements and repetitive sequences.Seurat 对象 ds_index 包含通过单细胞RNA测序（scRNA-seq）与单细胞T细胞受体测序（scTCR-seq），从1名在嵌合抗原受体T细胞治疗（CAR-T细胞治疗，Chimeric Antigen Receptor T-Cell Therapy, CAR-T）后出现帕金森综合征的索引患者的脑脊液（CSF）及外周血单个核细胞（PBMC）中获取的全部细胞。样本于多个时间点纵向采集（CAR-T治疗后第20天至第204天）。 T_cells_index 为ds_index的子集，包含从该索引患者的脑脊液与外周血单个核细胞混合样本中鉴定出的T细胞。 ds_control 包含通过单细胞RNA测序与单细胞T细胞受体测序，从以下外周血单个核细胞样本中获取的全部细胞： 1. 该索引患者的早期时间点样本（CAR-T治疗前及CAR-T治疗后第7天）； 2. 3名接受西达基奥仑赛（cilta-cel）治疗的对照患者的匹配时间点样本（CAR-T治疗前及CAR-T治疗后第9/10天）。 T_cells_control 为ds_control的子集，包含从该索引患者（CAR-T治疗前及CAR-T治疗后第7天）以及3名接受西达基奥仑赛（cilta-cel）治疗的对照患者（CAR-T治疗前及CAR-T治疗后第9/10天）的外周血单个核细胞样本中鉴定出的T细胞。 R 脚本 此外，本数据集附带3个R脚本： Script_index_QC_combination_of_index_datasets（索引患者质控脚本）：详述了针对索引患者数据集的质控、标准化与整合分析流程。 Script_index_analysis_ds_T_cell_subset（索引患者T细胞子集分析与可视化脚本）：包含用于生成索引患者数据集相关图表及执行下游分析的代码，涵盖克隆特征与表型表征分析。 Script_control_QC_and_downstream（早期时间点与对照队列处理脚本）：详述了针对索引患者及对照队列的早期时间点样本（ds_control与T_cells_control）的质控、Seurat对象构建及可视化步骤。 慢病毒整合分析方法概述（Lentiviral Integration Analysis） 数据预处理 接头去除与质量修剪 原始测序reads使用cutadapt工具处理以去除接头序列；随后使用Trimmomatic工具，以参数SLIDINGWINDOW:4:20和MINLEN:50进行进一步的低质量碱基修剪及长度过滤。 去除CAR构建序列reads 使用grep与seqtk工具识别并移除携带CAR构建序列的reads，以确保仅保留目标基因组序列用于后续分析。 比对与组装 基因组参考序列构建 将慢病毒序列追加至人类基因组参考序列（hg19）后，构建自定义基因组参考序列（hg19_lenti.fa）。 序列比对 使用高灵敏度比对工具BWA-MEM v0.7.17将修剪后的reads比对至自定义参考基因组；使用SAMtools v1.21转换并排序比对文件以用于下游处理。 嵌合reads识别 使用samtools工具提取跨越人类基因组与慢病毒整合位点的嵌合reads，并进行后续处理。 整合位点分析 整合位点注释 使用BEDTools v2.31识别与基因组特征存在重叠的显著整合位点；随后使用HOMER v5.1对已注释的整合位点进行进一步处理，将插入片段分类至不同基因组区域（如启动子、内含子、外显子、基因间区）。 可视化与结果分析 染色体分布 使用Python脚本绘制整合位点在各染色体（单倍体区域除外）上的分布可视化图。 特征富集分析 使用HOMER工具，通过观测值/期望值比值及log2富集评分，对基因组特征（如转座元件（Transposable Elements, TEs）、启动子）进行富集分析。 总结 本分析流程可从约100万个细胞中鉴定出约19046个独特的慢病毒整合位点；分析结果证实慢病毒载体偏好整合至具有转录活性的基因组区域，包括转录元件与重复序列。