Mining for viral fragments in methylation enriched sequencing data

Abstract

Most next generation sequencing experiments generate more data than is usable for the experimental set up. For example, methyl-CpG binding domain (MBD) affinity purification based sequencing is often used for DNA-methylation profiling, but up to 30% of the sequenced fragments cannot be mapped uniquely to the reference genome. Here we present and evaluate a methodology for the identification of viruses in these otherwise unused paired-end MBD-seq data. Viral detection is accomplished by mapping non-reference alignable reads to a comprehensive set of viral genomes. As viruses play an important role in epigenetics and cancer development, 92 (pre)malignant and benign samples, originating from two different collections of cervical samples and related cell lines, were used in this study. These samples include primary carcinomas (n = 22), low- and high-grade cervical intraepithelial neoplasia (CIN1 and CIN2/3 - n = 2/n = 30) and normal tissue (n = 20), as well as control samples (n = 17). Viruses that were detected include phages, adenoviruses, herpesviridae and HPV. HPV, which causes virtually all cervical cancers, was identified in 95% of the carcinomas, 100% of the CIN2/3 samples, both CIN1 samples and in 55% of the normal samples. Comparing the amount of mapped fragments on HPV for each HPV-infected sample yielded a significant difference between normal samples and carcinomas or CIN2/3 samples (adjusted p-values resp. <10−5, <10−5), reflecting different viral loads and/or methylation degrees in non-normal samples. Fragments originating from different HPV types could be distinguished and were independently validated by PCR-based assays in 71% of the detections. In conclusion, although limited by the a priori knowledge of viral reference genome sequences, the proposed methodology can provide a first confined but substantial insight into the presence, concentration and types of methylated viral sequences in MBD-seq data at low additional cost.

Genome-wide DNA Methylation Profiling Reveals Methylation Markers Associated with 3q Gain for Detection of Cervical Precancer and Cancer

“…silencing when affecting promoter regions, and an advantage over commonly used techniques detecting only methylation of single CpGs (32,33). Furthermore, using MBD-Seq, we could also detect methylated HPV DNA by analysis of the nonaligned reads (24). In the current study, both in vitro and patientobtained samples comprising all stages of cervical carcinogenesis were used, which allowed us to evaluate the succession of methylation events and the selection of methylation targets which are biologically relevant and associated with hrHPVmediated carcinogenesis.…”

“…For genome-wide DNA methylation profiling by MBD-Seq, various cell lines and frozen cervical tissue specimens were used. Cell lines consisted of (i) primary human foreskin keratinocytes (HFK) from two donors, (ii) HFK from one donor transduced with HPV16E6E7 at passage 7 and passage 30, repre-senting the preimmortal and immortal stage, respectively (17), (iii) consecutive passages of HPV16-(FK16A and FK16B) and HPV18-(FK18A and FK18B) immortalized keratinocytes, including early (passages [11][12][13][14][15][16][17][18][19][20][21][22][23][24][25][26][27][28][29][30], middle (passages 30-70), and late (>passage 70) immortal passages, and (iv) cervical cancer cell line SiHa. The tissue specimens comprised CIN1 (n ¼ 3), CIN2/3 (n ¼ 12), and SCC (n ¼ 10).…”

“…Genome-wide DNA methylation profiling was performed by MBD-Seq, using the MethylCap kit v2 (Diagenode) with 500 ng input DNA combined with Illumina Genome Analyzer IIx pairedend sequencing as described before (24,25). Next, Bowtie 1.0.0 was used to map the obtained paired-end reads from fastq-files to the human reference genome GRCh37 (26).…”

Computational epigenomics: challenges and opportunities

“…In this Research Topic, we collected a number of contributions in the field of computational epigenomics covering three main research areas: (i) computational analyses tackling important issues closely related to the experimental method used to generate epigenetic data (Flensburg et al, 2014 ; Ji et al, 2014 ; Mensaert et al, 2015 ), (ii) computational approaches useful to overcome pitfalls associated to the analysis of a given epigenetic layer (Barozzi et al, 2014 ; Cairns et al, 2014 ; Robinson et al, 2014 ), and (iii) studies on the integration of multiple epigenetic layers (de Pretis and Pelizzola, 2014 ; Fejes et al, 2014 ; Osella et al, 2014 ).…”

“…In some cases, a non-trivial subset of the DNA fragments sequenced in MBD-seq experiments, based on affinity purification through a methyl-CpG binding protein, could not be assigned to the expected reference genome. It was then shown how it is possible to assess this unanticipated proportion of unmapped reads to profile methylated viral sequences, which can be particularly relevant in certain studies (e.g., oncoviruses; Mensaert et al, 2015 ). On the other hand, reads from methylated DNA were shown to be over-represented in data from whole-genome bisulfite sequencing experiments.…”