Park, S.-H. | bioRxiv Science

bioRxiv2018

Integrative analysis of DNA methylation suggests down-regulation of oncogenic pathways and reduced de-novo mutation in survival outliers of glioblastoma

The study of survival outliers of glioblastoma (GBM) can have important implications on gliomagenesis as well as in the identification of ways to alter clinical course on this almost uniformly lethal cancer type. However, current studied epigenetic and genetic signatures of the GBM outliers have failed to identify unifying criteria to characterize this unique group of patients. In this study, we profiled the global DNA methylation pattern of mainly IDH1 wild type survival outliers of glioblastoma and performed comprehensive enrichment analyses with genomic and epigenomic signatures. We found that the genome of long-term survivors in glioblastoma is differentially methylated relative to short-term survivor patients depending on CpG density: hypermethylation near CpG islands (CGIs) and hypomethylation far from CGIs. Interestingly, these two patterns are associated with distinct oncogenic aspects in gliomagenesis. The hypomethylation pattern at the region distant from CGI is associated with lower rates of de novo mutations while the hypermethylation at CGIs correlates with transcriptional downregulation of genes involved in cancer progression pathways. These results extend our understanding of DNA methylation of survival outliers in glioblastoma in a genome-wide level, and provide insight on the potential impact of DNA hypomethylation in cancer genome.

cancer biology↗

bioRxiv2016

Contamination as a major factor in poor Illumina assembly of microbial isolate genomes

The nonhybrid hierarchical assembly of PacBio long reads is becoming the most preferred method for obtaining genomes for microbial isolates. On the other hand, among massive numbers of Illumina sequencing reads produced, there is a slim chance of re-evaluating failed microbial genome assembly (high contig number, large total contig size, and/or the presence of low-depth contigs). We generated Illumina-type test datasets with various levels of sequencing error, pretreatment (trimming and error correction), repetitive sequences, contamination, and ploidy from both simulated and real sequencing data and applied k-mer abundance analysis to quickly detect possible diagnostic signatures of poor assemblies. Contamination was the only factor leading to poor assemblies for the test dataset derived from haploid microbial genomes, resulting in an extraordinary peak within low-frequency k-mer range. When thirteen Illumina sequencing reads of microbes belonging to genera Bacillus or Paenibacillus from a single multiplexed run were subjected to a k-mer abundance analysis, all three samples leading to poor assemblies showed peculiar patterns of contamination. Read depth distribution along the contig length indicated that all problematic assemblies suffered from too many contigs with low average read coverage, where 1% to 15% of total reads were mapped to low-coverage contigs. We found that subsampling or filtering out reads having rare k-mers could efficiently remove low-level contaminants and greatly improve the de novo assemblies. An analysis of 16S rRNA genes recruited from reads or contigs and the application of read classification tools originally designed for metagenome analyses can help identify the source of a contamination. The unexpected presence of proteobacterial reads across multiple samples, which had no relevance to our lab environment, implies that such prevalent contamination might have occurred after the DNA preparation step, probably at the place where sequencing service was provided.

genomics↗