Transcriptomics and epigenomics datasets of primary brain cancers in formalin-fixed paraffin embedded format

Abstract

The access of public omics-based datasets is of paramount importance in brain cancer research as allows the proposal and validation of both biomarkers and therapeutic targets in gliomas, especially in the most prevalent and aggressive glioblastomas. Taking profit of current advances in next generation sequencing and DNA methylation profiling, we have created datasets from approximately 150 formalin-fixed paraffin embedded (FFPE) tumours. These datasets enable for the first time integrative transcriptional and epigenetics studies in a context that consider the degradation and fixation-derived chemical alterations of the most extended archiving format in hospitals, and provide an independent cohort from current public databases for further validation of putative novel biomarkers. Alongside with the most profusely known glioblastomas, astrocytomas and oligodendrogliomas, we have also included for comparison purposes few examples of rare tumours that are often neglected in brain cancer research. Taken together, we provide a valuable tool to explore combined gene expression and DNA methylation patterns in the study of gliomas and glioneuronal tumours.

Fig. 1

Overview of the multiomics datasets from FFPE-derived brain tumours. (a) Workflow and experimental design to obtain the transcriptomics and epigenomics data. (b) Final number of newly generated datasets, with the proportion of brain tumours subtypes.

Fig. 2

Transcriptional profiling quality details. (a) Concentration (left), RIN (middle) and DV200 values (right) of RNA extracted from tumour samples that were finally considered for sequencing. (b–e) Overview of the multiQC report illustrating: average Phred scores in all samples across all read bases (b) and across all counts (c); GC content per sequence (d); distribution of sequence length (e). (f) Summary of the key metrics related to alignment and fragment consistency as generated by the Salmon tool: ISF, inconsistent fragment start; ISR, inconsistent fragment strand; Assigned and Compatible fragments; Concordant and consistent mappings; Inconsistent/Orphan mappings; SF, single fragments; SR, single reads. The obtained values were the expected for high quality datasets.

Fig. 3

Methylation profiling quality details. (a) Concentration (left) and average ΔC_T values (right) of DNA extracted from tumour samples that were finally considered for beadchips hybridization. (b) Profile of raw and normalized beta values from all filtered samples. c, Distribution of CpGs that passed the filters across genomic locations related to CpG islands, in which shores, shelves and open seas are regions <2 kb, between >2 and <4 kb, and >4 kb from CpG islands, respectively.