Bcl2fastq program

Base calls from sequencing were converted into fastq format using Illumina’s bcl2fastq program and demultiplexed by the University of Washington sequencing core. Reads were adaptor-clipped using trim_galore with default settings. Trimmed reads were mapped to the human reference genome (GRch38) using the STAR program (version 2.5.2b). Uniquely mapping reads were extracted, and duplicates were removed based on unique molecular identifier (UMI) sequences. To generate expression matrices, the number of UMIs for each cell mapping to the exonic and intronic regions of each gene was calculated. Potential ambient RNA reads were estimated and removed using the R package SoupX [38 (link)]. Doublets were then identified using the Python package Scrublet [39 (link)]. Further analysis for quality filtering was performed using the Seurat R package (version 3.2.2). Only cells with a total read count <10,000 and number of genes detected >100 were kept. To remove potential dead cells from the analysis, cells with >15% mitochondrial reads were filtered out. In total, we obtained 82,133 cells from 4 conditions.

Sequencing data were demultiplexed using Illumina bcl2fastq© program. Demultiplexed paired FASTQ sequences were imported in QIIME2 artifact format and analysed with QIIME2 v2020.8. Used workflow is described in detail in the following link: https://github.com/Marylou8/Metataxonic-analysis-using-Qiime2-workflow. Quality control was carried out using the DADA2 pipeline [33 (link)] incorporated into QIIME2 [34 , 35 (link)]. The DADA2 program filtered out PhiX reads, removed chimeric sequences and assigned reads into Amplicon Sequence Variants (ASVs). Taxonomic annotation for bacteria was obtained using SILVA v138 database [36 (link)]. Taxonomic annotation for fungi was obtained using UNITE v8.2 2020 database [37 ]. Chloroplast and mitochondrial contaminants were detected and filtered using the QIIME2 “taxa filter-table” and “taxa filter-seqs” commands.
To filter out low-abundance features, we follow the approach of Morton (https://forum.qiime2.org/t/ancom-giving-strange-w-values/1002/11) where those features which do not sum at least 10 sequences among all samples, as well as those that only appear in one sample were filter out (command “feature-table filter-features”). Differential abundance analysis was analysed with ANCOM test [38 ]. The data-set will be submitted to National Center for Biotechnology Information (NCBI).

The sequencing results were demultiplexed and converted to FASTQ format using the Illumina bcl2fastq program (version 2.20). The Cell Ranger pipeline (https//support.10xgenomics.com/single-cell-geneexpression/software/pipelines/latest/what-is-cell-ranger, version 6.1.1) was used for demultiplexing, barcode processing, and single-cell 3’ gene counting. The scRNA-seq data were aligned to the Ensembl GRCh38/GRCm38 reference genome.10X Genomics Chromium Single-Cell 3’Solution was used to process a total of 110,000 single cells obtained from sorted CAR-T cells and apheresis T cells from six patients. Seurat (version 3.1.1) was used to load the Cell Ranger output to perform dimensional reduction, clustering and analysis of the scRNA-seq data. A total of 60,000 cells met quality control criteria. All genes expressed in fewer than three cells were excluded. The number of genes expressed per cell was considered low if greater than 500 and high if less than 5000. UMI counts below 500 were not considered, and gene expression derived from mitochondrial DNA was less than 25%.