The blood DNA virome in 8,000 humans

Abstract

The characterization of the blood virome is important for the safety of blood-derived transfusion products, and for the identification of emerging pathogens. We explored non-human sequence data from whole-genome sequencing of blood from 8,240 individuals, none of whom were ascertained for any infectious disease. Viral sequences were extracted from the pool of sequence reads that did not map to the human reference genome. Analyses sifted through close to 1 Petabyte of sequence data and performed 0.5 trillion similarity searches. With a lower bound for identification of 2 viral genomes/100,000 cells, we mapped sequences to 94 different viruses, including sequences from 19 human DNA viruses, proviruses and RNA viruses (herpesviruses, anelloviruses, papillomaviruses, three polyomaviruses, adenovirus, HIV, HTLV, hepatitis B, hepatitis C, parvovirus B19, and influenza virus) in 42% of the study participants. Of possible relevance to transfusion medicine, we identified Merkel cell polyomavirus in 49 individuals, papillomavirus in blood of 13 individuals, parvovirus B19 in 6 individuals, and the presence of herpesvirus 8 in 3 individuals. The presence of DNA sequences from two RNA viruses was unexpected: Hepatitis C virus is revealing of an integration event, while the influenza virus sequence resulted from immunization with a DNA vaccine. Age, sex and ancestry contributed significantly to the prevalence of infection. The remaining 75 viruses mostly reflect extensive contamination of commercial reagents and from the environment. These technical problems represent a major challenge for the identification of novel human pathogens. Increasing availability of human whole-genome sequences will contribute substantial amounts of data on the composition of the normal and pathogenic human blood virome. Distinguishing contaminants from real human viruses is challenging.

Fig 1. Study design.

The flowchart summarizes the steps followed to identify viral content in the human blood DNA from whole-genome sequencing reads.

Fig 2. Viral content.

The heatmap shows the presence of reads of viral nature in sequencing reactions of blood from 8,240 individuals. Extensive phage and other viral DNA is found in sequencing reactions, but it is almost universally associated to including phiX174 phage spike-in in the reaction (used in 60% of samples). For reference, we include the ubiquitous identification of human endogenous retrovirus (HERVs) sequences in the pool of unmapped reads.

Fig 3. Prevalence and abundance of human DNA viruses and retroviruses in 8,240 individuals.

A. Frequency of 19 human viruses in the study population ranked according to their prevalence. B. The viral load of human viruses represented on the x-axis as genome copies per 100,000 human cells; the bar represents the median.

Fig 4. Genome coverage of selected human viruses.

Shown are the alignment of reads contributed by all individuals carrying the corresponding virus. The depth of coverage (y-axis) changes in scale as a reflection of the viral abundance and prevalence. Gaps in coverage (e.g., in EBV) generally reflect repetitive regions that are masked during data processing.

Fig 5. Integration of human herpesvirus 6.

The two populations of HHV6A andHHV6B are present in a bimodal distribution. The frequency of integrated viruses, at approximately 0.5 per cell corresponds to the haploid nature of the integration in the case of inherited, vertical transmission—from one of the parents. The identification of chimeric reads, or paired human-virus reads is shown for a substantial proportion of integrated HHV6 (green dots). The bar represents the median.

Fig 6. Relative proportion and viral load in the context of age, sex and ancestry.

The relative proportion, normalized to 100% for visualization purposes (A, C and E) and distribution of observed viral loads (B, D and F) are depicted for the 8 viruses that have the largest prevalence in the study. Among the 4,505 with demographic information, the ancestries were: EUR, European = 3,048; AFR, African = 665; MDE, Middle Eastern = 94; EAS, East Asian = 91; CSA, Central South Asian = 54; AMR, Admixed American = 8; Multi-Racial and Others = 545.