An integrated single-cell transcriptomic dataset for non-small cell lung cancer

As single-cell RNA sequencing (scRNA-seq) has emerged as a great tool for studying cellular heterogeneity within the past decade, the number of available scRNA-seq datasets also rapidly increased. However, reuse of such data is often problematic due to a small cohort size, limited cell types, and insufficient information on cell type classification. Here, we present a large integrated scRNA-seq dataset containing 224,611 cells from human primary non-small cell lung cancer (NSCLC) tumors. Using publicly available resources, we pre-processed and integrated seven independent scRNA-seq datasets using an anchor-based approach, with five datasets utilized as reference and the remaining two, as validation. We created two levels of annotation based on cell type-specific markers conserved across the datasets. To demonstrate usability of the integrated dataset, we created annotation predictions for the two validation datasets using our integrated reference. Additionally, we conducted a trajectory analysis on subsets of T cells and lung cancer cells. This integrated data may serve as a resource for studying NSCLC transcriptome at the single cell level.