scSpAS: an integrated single-cell and spatial atlas of transcript isoform regulation

Alternative transcript initiation, splicing, and 3'-end processing jointly shape transcript isoform diversity, but these regulatory layers remain fragmented across separate single-cell and spatial resources. Here, we present scSpAS, an integrated database of alternative transcription start site (ATS), alternative splicing, full-length transcript isoforms, and alternative polyadenylation (APA) in human and mouse. It contains 123 datasets, including over 418 000 cells and 82 000 spatial spots across healthy and disease contexts, covering 13 cancer types, 5 infectious conditions, 8 inflammatory diseases, and 90 healthy samples. Based on uniformly standardized pipelines, we utilized 5' capture and 3' capture single-cell RNA-sequencing data to identify ATS and APA sites, and the Smart-seq2 and long-read sequencing data to identify the alternative splicing events and isoforms. In total, scSpAS contains >648 000 alternative splicing events, 432 000 ATS, and 621 000 APA sites, as well as over 551 000 transcript isoforms. The database further integrates regulatory annotations derived from CLIP-seq, ATAC-seq, and CAGE-seq, together with predicted splicing regulatory motifs. scSpAS enables users to retrieve events, isoforms and genes, examine differential usage at cell-type and spatial levels, and in different biological and physiological conditions; furthermore, users can inspect the potential regulatory mechanisms of all events. It is freely available at https://scspas.chenlulab.com/.