Transcriptome complexity and regulation conferred by genomic structural variations

Integration of structural variations (SVs) with transcriptome and methylome within a given cancer cell line shows promise for elucidating the role of these aberrations in perturbing biological mechanisms. Here, on the basis of the model leukemia cell line U937 that harbors the MLLT10-PICALM and PICALM-MLLT10 fusion genes formed by translocation, we used multiple platforms to generate layered data including genome (ONT ultra-long- and Illumina reads), transcriptome (PacBio HiFi- and ONT full-length reads, and Illumina reads) and methylome (profiled by ultra-long reads). Leveraging the ultra-long whole-genome sequencing (WGS) data, we provided a panoramic view of the U937 genome through assembling the highly contiguous genome de novo, phasing major histocompatibility complex (MHC) haplotypes, measuring telomere repeat length, resolving heteroplasmy patterns with the complete mitochondrial sequences, and detecting the full spectrum of SVs. Depending on long- and short-read RNA-seq data, we projected the SV events onto the involved transcript sequences, unveiling the modified transcript structure and gene dosage. Taking advantage of the ultra-long reads simultaneously traversing the MLLT10-PICALM promoter and breakpoint junction, we unambiguously identified allele-specific CpG methylation states, disclosing that the expression of the wild-type MLLT10 gene is abrogated by hypermethylation of its promoter. Collectively, this integrative analysis provides a general framework to investigate the complicated effects of SVs on transcriptome complexity and regulation in a personalized genomic context.